Technical Program, CA Bouman, GG Marcu

Tags: stereoscopic displays, Stereoscopic Displays and Applications, horizontal disparity, stereoscopic 3D, field of view, application, South Korea, Japan, applications, autostereoscopic displays, autostereoscopic display, Telecommunications Research Institute, stereoscopic display
Content: IS&T/SPIE 18th Annual Symposium
Technical Program
15-19 January 2006 San Jose Marriott and San Jose Convention Center · San Jose, California USA Sponsored by:
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
1
Technical Program IS&T/SPIE 18th Annual Symposium
15­19 January 2006 San Jose Marriott and San Jose Convention Center · San Jose, California USA Conferences · Continuing Education · Technical Exhibition · Demonstrations
IS&T/SPIE would like to express their deepest appreciation to the symposium chairs, conference chairs, and program committees who have so generously given of their time and advice to make this symposium possible. The symposium, like our other conferences and activities, would not be possible without the dedicated contribution of our participants and members.
Welcome We are all too busy. If your jobs are like ours, then you are expected to produce more innovation, more products, more research, more students, more money, but all with less time. So how does participating in EI 2006 fit into this picture? Well, here is the straight answer as to why the Electronic Imaging Symposium is a great investment of your time and resources. · Defines the cutting edge in imaging research - The EI symposium, and its associated publication the Journal of Electronic Imaging, have been defining the envelope of high impact digital imaging research since their inception. Because of its leading role, EI is where the innovation leaders in imaging research go to find out about the newest imaging systems, methods, instrumentation, and algorithms. · Provides broad coverage with tremendous technical depth - The unique structure of EI leads to a rich marketplace of ideas where a broad range of topics are discussed in great technical depth. Each year, the EI symposium starts new conferences on important new topics and reinvigorates highly respected existing conferences through the incorporation of new research and participants. · Promotes an atmosphere for professional networking - Because of its organization into primarily single-tracked conferences, the symposium has a very personal and friendly atmosphere that gives technical experts and researchers an opportunity to learn what they need and also, to efficiently network on a professional level. · Enhances your professional value - EI gives you great opportunities to publish high-quality research, meet widely recognized experts in your field, and keep up on "hot" technical trends. All these things enhance the value of both you and your organization. · Provides opportunities for professional participation - Each year new people get involved and infuse EI with fresh new ideas. There are numerous opportunities to participate, and we want both your input and your participation! If you have ideas, feel free to speak to your conference chair, or other EI representatives about getting involved. · Leverages the Silicon Valley community - Every year EI is held in San Jose, the heart of Silicon Valley and the hub of the international information technology industry. Each year we look forward to visiting the institutions that are shaping the future of the imaging industry, and we also don't mind the warm weather! We look forward to seeing you both this year and for years to come.
2006 Symposium Chairs Charles A. Bouman, Purdue Univ.
Gabriel G. Marcu, Apple Computer, Inc.
2
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Electronic Imaging Contents Special Events . . . . . . . . . . . . . . 6 Plenaries . . . . . . . . . . . . . . . . . . 7 Conferences: Conference Index . . . . . . . . . 3 Daily Overview . . . . . . . . . . . . 4 Technical Conferences . 10-67 Course Daily Overview . . . . . . . 5 General Information . . . . . . . 8-9 Technical Abstract Summaries . . . . . . . . . 68-267 Participants . . . . . . . . . . 268-278 Publication Order form . . . . . 279 Proceedings . . . . . . . . . . . . . 280
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
1
Electronic Imaging
Symposium Chairs Charles A. Bouman, Purdue Univ. Gabriel G. Marcu, Apple Computer, Inc. Symposium Organizing Committee Thrasyvoulos N. Pappas, Northwestern Univ. Andrew J. Woods, Ctr. for Marine Science and Technology/Curtin Univ. of Technology (Australia) Short Course Chairs Edward J. Delp III, Purdue Univ. Michael A. Kriss, Consultant Exhibition Chair A. Ufuk Agar, Garanti Technologies (Turkey) Demonstration Chair Neil A. Dodgson, Univ. of Cambridge (United Kingdom)
Technical Organizing Committee Jan P. Allebach, Purdue Univ. John G. Apostolopoulos, Hewlett-Packard Labs. Jaakko T. Astola, Tampereen Teknillinen Yliopisto (Finland) Morley M. Blouke, Ball Aerospace & Technologies Corp. Mark T. Bolas, Univ. of Southern California Katy Bцrner, Indiana Univ. Charles A. Bouman, Purdue Univ. Surendar Chandra, Univ. of Notre Dame Edward Y. Chang, Univ. of California/Santa Barbara Hui Chao, Hewlett-Packard Co. Chang Wen Chen, Florida Institute of Technology Brian D. Corner, U.S. Army Natick Soldier Ctr. Reiner Creutzburg, Fachhochschule Brandenburg (Germany) Luke C. Cui, Lexmark International, Inc. Scott J. Daly, Sharp Labs. of America, Inc. Edward J. Delp III, Purdue Univ. Jeffrey M. DiCarlo, HewlettPackard Labs. Neil A. Dodgson, Univ. of Cambridge (United Kingdom) Edward R. Dougherty, Texas A&M Univ. Karen O. Egiazarian, Tampereen Teknillinen Yliopisto (Finland) Robert F. Erbacher, Utah State Univ. Reiner Eschbach, Xerox Corp. Theo Gevers, Univ. van Amsterdam (Netherlands) Carsten Griwodz, Univ. of Oslo (Norway) Matti T. Grцhn, CSC-Scientific Computing Ltd. (Finland) Alan Hanjalic, Technische Univ. Delft (Netherlands) Francisco Hideki Imai, Rochester Institute of Technology Nasser Kehtarnavaz, The Univ. of Texas at Dallas Phillip A. Laplante, The Pennsylvania State Univ. Longin Jan Latecki, Temple Univ. Peng Li, GEO-Centers, Inc.
Xiaofan Lin, Hewlett-Packard Labs. Gabriel G. Marcu, Apple Computer, Inc. Russel A. Martin, Foveon USA Ian E. McDowall, Fakespace Labs., Inc. Fabrice Meriaudeau, Univ. de Bourgogne (France) John O. Merritt, The Merritt Group Eric L. Miller, Northeastern Univ. Yoichi Miyake, Chiba Univ. (Japan) David M. Mount, Univ. of Maryland/College Park Nasser M. Nasrabadi, Army Research Lab. Kurt Niel, Fachhochschule Wels (Austria) Thrasyvoulos N. Pappas, Northwestern Univ. Syed A. Rizvi, CUNY/College of Staten Island Jonathan C. Roberts, Univ. of Kent (United Kingdom) Bernice E. Rogowitz, IBM Thomas J. Watson Research Ctr. Mitchell Rosen, Rochester Institute of Technology Amir Said, Hewlett-Packard Labs. Nitin Sampat, Rochester Institute of Technology Simone Santini, Univ. of California/San Diego Raimondo Schettini, DISCo/ Univ. degli Studi di MilanoBicocca (Italy) Nicu Sebe, Univ. van Amsterdam (Netherlands) Kazem Taghva, Univ. of Nevada/Las Vegas Jarmo H. Takala, Tampereen Teknillinen Yliopisto (Finland) Matthew Tocheri, Arizona State Univ. Shoji Tominaga, Osaka ElectroCommunication Univ. (Japan) Rudolf L. van Renesse, VanRenesse Consulting (Netherlands) Ping Wah Wong, IDzap LLC Andrew J. Woods, Ctr. for Marine Science and Technology/Curtin Univ. of Technology (Australia) Angela Y. Wu, American Univ.
2
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference Index
Program on
3D Imaging, Interaction, and Measurement
Mon-Weds 6055A Stereoscopic Displays and Applications XVII (Woods, Dodgson, Merritt) . . . . . . . . . . . . . . . . . . . . . . . p. 10
Thurs
6055B The Engineering Reality of Virtual Reality 2006 (Bolas, McDowall) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 14
Mon-Tues 6056 Three-Dimensional Image Capture and Applications VI (Corner, Li, Tocheri) . . . . . . . . . . . . . . . . . . . . . . . . . . p. 15
Program on Imaging, Visualization, and Perception Mon-Wed 6057 Human Vision and Electronic Imaging XI (Rogowitz, Pappas, Daly) . . . . . . . . . . . . . . . . . . . . . . . . p. 17 Tues-Thurs 6058 Color Imaging XI: Processing, Hardcopy, and Applications (Eschbach, Marcu) . . . . . . . . . . . . . . . . . . p. 20 Tues-Thurs 6059 Image Quality and System Performance III (Cui, Miyake) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 22 Mon-Tues 6060 Visualization and Data Analysis 2006 (Erbacher, Roberts, Grцhn, Bцrner) . . . . . . . . . . . . . . . . . . . . . . . . . p. 24 Tues-Thurs 6061 Internet Imaging VII (Gevers, Santini, Schettini) . p. 26 Mon-Tues 6062 Spectral Imaging: Eighth International Symposium on Multispectral Color Science (Rosen, Imai, Tominaga) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 28
Program on Image Processing Mon-Tues 6063 Real-Time Image Processing III (Kehtarnavaz, Laplante) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 30 Mon-Tues 6064A Image Processing: Algorithms and Systems V (Dougherty, Astola, Egiazarian) . . . . . . . . . . . . . . . . . . . . p. 32 Tues-Weds 6064B Applications of Neural Networks and Machine Learning in Image Processing X (Nasrabadi, Rizvi) . . . p. 34 Mon-Weds 6065 Computational Imaging IV (Bouman, Miller, Pollak) . . . . . . . . . . . . . . . . . . . . . . . . . p. 36 Tues-Weds 6066 Vision Geometry XIV (Latecki, Mount, Wu) . . . . . p. 39 Weds-Thurs 6067 Document Recognition and Retrieval XIII (Taghva, Lin) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 41
Program on Digital Imaging Sensors and Applications Tues-Thurs 6068 Sensors, Cameras, and Systems for Scientific/ Industrial Applications VIII (Blouke) . . . . . . . . . . . . . . . p. 43 Mon-Tues 6069 Digital Photography II (Sampat, DiCarlo, Martin) p. 45 Mon-Tues 6070 Machine Vision Applications in Industrial Inspection XIV (Meriaudeau, Niel) . . . . . . . . . . . . . . . . . p. 47
Program on Multimedia Processing and Applications Weds-Thurs 6071 Multimedia Computing and Networking 2006 (Chandra, Griwodz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 49 Mon-Thurs 6072 Security, Steganography, and Watermarking of Multimedia Contents VIII (Delp, Wong) . . . . . . . . . . . . p. 51 Tues-Thurs 6073 Multimedia Content Analysis, Management, and Retrieval 2006 (Chang, Hanjalic, Sebe) . . . . . . . . . . . . . p. 55 Mon-Tues 6074 Multimedia on Mobile Devices II (Creutzburg, Takala, Chen) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 58 Mon-Tues 6076 Digital Publishing (Allebach, Chao) . . . . . . . . . . . p. 62 Program on Optical Security and Anti-Counterfeiting Tues-Thurs 6075 Optical Security and Counterfeit Deterrence Techniques VI (van Renesse) . . . . . . . . . . . . . . . . . . . . . p. 60 Program on Visual Communications and Image Processing Tues-Thurs 6077 Visual Communications and Image Processing 2006 (Apostolopoulos, Said) . . . . . . . . . . . . . . . . . . . . . . . . . . p. 64
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
3
Conference Daily Overview
Monday
Tuesday
Wednesday
Thursday
Program on 3D Imaging, Interaction, and Measurement Conf. 6055A Stereoscopic Displays and Applications XVII (Woods, Dodgson, Merritt), p. 10 Conf. 6056 Three-Dimensional Image Capture and Applications VI (Corner, Li, Tocheri), p. 15
Conf. 6055B The Engineering Reality of Virtual Reality 2006 (Bolas, McDowall), p. 14
Program on Imaging, Visualization, and Perception Conf. 6057 Human Vision and Electronic Imaging XI (Rogowitz, Pappas, Daly), p. 17 Conf. 6058 Color Imaging XI: Processing, Hardcopy, and Applications (Eschbach, Marcu), p. 20 Conf. 6059 Image Quality and System Performance III (Cui, Miyake), p. 22 Conf. 6061 Internet Imaging VII (Gevers, Santini, Schettini), p. 26 Conf. 6060 Visualization and Data Analysis 2006 (Erbacher, Roberts, Grцhn, Bцrner), p. 24 Conf. 6062 Spectral Imaging: Eighth International Symposium on Multispectral Color Science (Rosen, Imai, Tominaga), p. 28
Program on Image Processing
Conf. 6063 Real-Time Image Processing III (Kehtarnavaz, Laplante), p. 30 Conf. 6064A Image Processing: Algorithms and Systems V (Dougherty, Astola, Egiazarian), p. 32
Conf. 6064B Applications of Neural Networks and Machine Learning in Image Processing X (Nasrabadi, Rizvi), p. 34
Conf. 6065 Computational Imaging IV (Bouman, Miller, Pollak), p. 36
Conf. 6066 Vision Geometry XIV (Latecki, Mount, Wu), p. 39
Program on Digital Imaging, Sensors, and Applications
Conf. 6067 Document Recognition and Retrieval XIII (Taghva, Lin), p. 41
Conf. 6068 Sensors, Cameras, and Systems for Scientific/Industrial Applications VIII (Blouke), p. 43
Conf. 6069 Digital Photography II (Sampat, DiCarlo, Martin), p. 45
Conf. 6070 Machine Vision Applications in Industrial Inspection XIV (Meriaudeau, Niel), p. 47
Program on Multimedia Processing and Applications
Conf. 6071 Multimedia Computing and Networking 2006 (Chandra, Griwodz), p.49
Conf. 6072 Security, Steganography, and Watermarking of Multimedia Contents VIII (Delp, Wong), p. 51
Conf. 6073 Multimedia Content Analysis, Management, and Retrieval 2006 (Chang, Hanjalic, Sebe), p. 55
Conf. 6074 Multimedia on Mobile Devices II (Creutzburg, Takala, Chen), p. 58
Conf. 6076 Digital Publishing (Allebach, Chao), p. 62
Program on Optical Security and Anti-Counterfeiting Conf. 6075 Optical Security and Counterfeit Deterrence Techniques VI (van Renesse), p. 60 Program on Visual Communications and Image Processing Conf. 6077 Visual Communications and Image Processing 2006 (Apostolopoulos, Said), p. 64
4
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Sunday
Monday
Tuesday
Course Daily Overview
Wednesday
Thursday
Digital Imaging
SC753 Image Quality Evaluation for Digital Cameras Based on Existing ISO Standards (Wueller, Matherson) 8:30 am to 5:30 pm SC504 Introduction to CCD and CMOS Imaging Sensors and Applications (Janesick) 8:30 am to 5:30 pm
SC760 CCD Technology/ Digital Photographic Systems Technology (Theuwissen) 8:30 am to 5:30 pm SC762 Device Simulation for Image Quality Evaluation (Farrell) 8:30 am to 5:30 pm SC772 High Dynamic Range Techniques: From Acquisition to Display (Heidrich, Hoefflinger, Myszkowski) 8:30 am to 5:30 pm
SC513 Practical MTF and Noise performance measurement for Digital Cameras and Scanners (Burns, Williams) 8:30 am to 5:30 pm
Electronic Imaging Technology
Students: Save 50% on course registrations! Register for Courses at the registration desk.
SC066 Fundamentals of Electronic Image Processing (Weeks) 8:30 am to 5:30 pm, $450 / $53 SC589 Video Compression: Standards and Trends (Rabbani) 8:30 am to 5:30 pm
SC590 Advanced Digital Image and Video Enhancement Algorithms (Rabbani) 8:30 am to 5:30 pm SC763 Subband/Wavelet Scalable Video Coding (Woods) 8:30 am to 12:30 pm, $220 / $260
Electronic Imaging Applications
SC084 An Introduction to Cryptography and Digital Watermarking with Applications to Multimedia Systems and Forensics (Delp, Dittmann) 8:30 am to 5:30 pm SC060 Stereoscopic Display Application Issues (Merritt, Woods) 8:30 am to 5:30 pm SC766 Video Surveillance (Ebrahimi, Dufaux) 8:30 am to 5:30 pm
SC764 Filter Banks and Wavelets: Design and Use in Perceptual Coding (Schuller) 1:30 to 5:30 pm
SC761 Novel Spatially Adaptive Anisotropic Local Approximation Techniques in Image Processing (Katkovnik, Egiazarian, Astola) 8:30 am to 12:30 pm SC685 Content-based Image and Video Retrieval (Gevers, Sebe) 1:30 to 5:30 pm SC767 Practical Implementations of Machine Vision Systems within Technical Processes (Niel) 8:30 am to 12:30 pm
SC491 Neural Networks Applications in Image Processing (Nasrabadi) 8:30 am to 5:30 pm
Biometrics and Security
SC686 Biometrics: Appli-cations, Technologies, Standards and Evaluation (Vielhauer) 8:30 am to 5:30 pm SC759 Ownership of the Electronic Image: Overview of Copyright, Privacy, and Liability Issues (Williams) 1:30 to 5:30 pm
SC087 Optical Document Security (van Renesse) 8:30 am to 12:30 pm
Color and Perception
SC754 Human Shape Perception (Pizlo, Latecki) 8:30 am to 5:30 pm
SC765 The Biology, Physics, and Metrology of Color Perception (Appell) 8:30 am to 5:30 pm
SC516 Color Considerations for Liquid Crystal Displays (Marcu) 8:30 am to 12:30 pm
SC075 Effective Color Computing (Marcu) 1:30 to 5:30 pm
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
5
Special Events
Technical Group Meeting Members and nonmembers alike are invited to attend this informative meeting that provides excellent networking opportunities. Electronic Imaging Marriott Ballroom 3 Monday 16 January . . . . . . . . . . . . . . . . . . . . . . . . 7:30 to 9:30 pm Chair: Gabriel Marcu, Apple Computer, Inc. This group addresses diverse research, engineering, and specialized applications of electronic imaging devices or systems. Because of the diverse topical areas within electronic imaging, the technical group covers image processing, image capture, display and hardcopy, system integration and visualization. Application areas are just as far-reaching. They include industrial automation, graphic arts, aerospace sensing, remote sensing, document processing, highresolution television, medical imaging, and all areas of digital image processing, including analysis, compression and restoration. The group members are strongly encouraged to propose topics of interest for the next meeting and to submit short articles for publications in the Electronic Imaging Newsletter which serves to promote the topics of interest to the Group. About the IS&T/SPIE Electronic Imaging Technical Group Joint sponsorship by the IS&T and SPIE provides even more benefits and contacts for members of the technical group through the close partnership of the two societies. Both IS&T and SPIE members may join the technical group at the member rate of just $15 per year. Technical group benefits include: · a twice-yearly newsletter covering events in the field · an annual directory of members · discounts on selected publications, including the SPIE/IS&T copublished quarterly Journal of Electronic Imaging. Poster Session San Jose Convention Center, Exhibit Hall 1 Tuesday, 17 January . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Conference attendees are invited to the poster session. Authors of poster papers will be on hand during this session to answer questions and provide in-depth discussion concerning their papers. Attendees are requested to wear their conference registration badges to the poster session. Authors can set up posters after 10:00 am on Tuesday. Poster supplies (pushpins) will be available. Other supplies can be obtained from the Speakers' Audio Visual Desk. Posters can be previewed during the day of the event before the formal poster session begins at 5:30 pm. Authors must remove their papers at the conclusion of the poster reception. It is the author's responsibility to remove their posters immediately after the session. Papers not removed will be considered unwanted and will be discarded. The Societies assume no responsibility for posters left up after the end of the poster reception.
All-Conference Reception Marriott Ballroom 4-6 Wednesday 18 January . . . . . . . . . . . . . . . . . . . . . 7:30 pm to 9:30 pm Plan to join us for this great opportunity to get to know your Electronic Imaging colleagues. All attendees are invited to relax and enjoy a pleasant evening with friends old and new! Symposium Demonstration Session San Jose Convention Center, Exhibit Hall 1 Tuesday 17 January . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm Conference Chair: Neil A. Dodgson, Univ. of Cambridge (United Kingdom) The highly-successful, interactive, hands-on demonstration of stereoscopic hardware, software, and display--traditionally a component of the Stereoscopic Display and Applications Conference--will be expanded this year to include research demonstrations and products related to the entire Electronic Imaging Symposium. 3D Phantogram Exhibit San Jose Convention Center, Concourse 1 Lobby Tuesday 17 January . . . . . . . . . . . . . . . . . . . . . . 10:00 am to 8:30 pm Wednesday 18 January . . . . . . . . . . . . . . . . . . . . 10:00 am to 4:00 pm Phantograms are a relatively new "3D art form" which place realistic three-dimensional images within hand's reach of the observer. In some ways like a hologram, but in other ways not, phantograms use conventional stereoscopic display technology in a special way to present images that are enchanting. Be prepared to experience a new reality with this innovative combination of art and technology. Exhibition and Bookfair Hours Tuesday 17 January . . . . . . . . . . . . . . . . . . . . . . 10:00 am to 5:00 pm Select exhibitors may choose to stay open to 8:30 pm during the Demonstration and Poster Sessions Wednesday 18 January . . . . . . . . . . . . . . . . . . . . 10:00 am to 4:00 pm 2006 Exhibitors Ukrainian Chapter of SPIE--The International Society for Optical Engineering 3D Consortium Photron USA Institute for Microelectronics Stuttgart ABBYY USA Software House Photonics Spectra Vision Systems Design Wiley Morgan Kaufman/Elsevier Vision Gates 360°
6
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Plenaries
Plenaries Marriott Ballroom 1-6 Tuesday 17 January . . . . . . . . . . . . . . . . . . . . . . . . . . 8:30 to 9:15 am Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at UrbanaChampaign, Recipient of the 2006 Electronic Imaging Scientist of the Year Award We live in a complex world. To solve complex problems, we need to combine knowledge and technologies from diverse fields. Image processing often plays a key role in these interdisciplinary problems. In a narrow sense, image processing comprises three areas: coding, enhancement/restoration/reconstruction, and analysis (mensuration/detection/recognition). These three areas are, of course, intimately related to each other. Many 2D images are perspective views of 3D objects and scenes. When we try to relate a 2D image to its originating 3D objects/scene, we enter the realm of computer vision. Computer vision techniques are increasingly being used in computer graphics and animation. One may take the position that in a broad sense, image processing subsumes computer vision and computer graphics. Finally, to solve many important problems, it may be advantageous, or necessary to use multimodal (especially, audio and visual) information. In this talk, we shall give two examples of interconnections. First: very low bitrate video coding using a 3D modelbased approach, which combines computer vision and computer graphics. Second: audio-visual speech recognition, which combines the audio and the visual modalities. Biography: Thomas S. Huang received his BS from National Taiwan University, and SM and SC.D. from the Massachusetts Institute of Technology, all in Electrical Engineering. He was on the Faculties of MIT and Purdue University before joining the University of Illinois at UrbanaChampaign in 1980, where he is currently William L. Everitt Distinguished Professor of Electrical and Computer Engineering, Professor of the Center for Advanced Study, Research Professor at Coordinated Science Laboratories, and Cochair of the Human Computer Intelligent Interaction major research theme at the Beckman Institute for Advanced Science and Technology. Huang's research interests lie in the broad area of Information Technology, but especially Multidimensional and Multimodal Signal Processing, with applications to human computer interaction, and multimedia data indexing, retrieval, and mining. He has published 21 books and more than 600 journal and conference papers in 2D digital filtering, digital holography, image and video compression, multimodal human computer interfaces, and multimedia data retrieval. He is a member of the National Academy of Engineering, a Fellow of SPIE, OSA, IAPR, and IEEE; and has received numerous awards, including: IEEE Jack S. Kilby Signal Processing Medal (co-recipient with Arun Netravali), and the King-Sun Fu Prize of the Int'l. Association of Pattern Recognition.
Wednesday 18 January . . . . . . . . . . . . . . . . . . . . . . . . 8:30 to 9:15 am Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard M. Leahy, Univ. of Southern California The combined revolutionary advances in recent years in molecular biology, imaging science and computing power make this a golden age for biomedical imaging. Central to modern biomedical imaging is the requirement for new computational imaging methods for noninvasive studies of human and animal anatomy and function. Computational tools are needed both for optimizing resolution and noise properties in reconstructed images and for the interpretation and statistical analysis of these images across modalites, subjects, and populations. I will describe some recent work on image formation and analysis in functional brain mapping and molecular imaging, and attempt to highlight common themes and open research questions relevant to computational imaging. Biography: Richard M. Leahy received the B.Sc. and Ph.D. degrees in Electrical Engineering from the University of Newcastle upon Tyne, England. In 1985 he joined the University of Southern California where he is a Professor of Electrical Engineering, Radiology and Biomedical Engineering and was Director of the Signal and Image Processing Institute from 1997 until 2003. He was General Chair of the 2004 IEEE International Symposium on Biomedical Imaging and is a Fellow of the IEEE. His research interests lie in the application of signal and image processing theory to anatomical and functional imaging with applications in neuroimaging, oncology, and gene expression.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
7
General Information
Electronic Imaging 2006 San Jose Convention Center 408 S. Almaden Boulevard, San Jose, CA 95110 San Jose Marriott Hotel 301 S. Market Street, San Jose, CA 95113 Registration Location and Information Hours San Jose Convention Center, Concourse 1 Lobby Courses Only: Sunday 15 January . . . . . . . . . . . . . . . . . . . . . . . 7:00 am to 10:00 am Conference Registration: Sunday 15 January . . . . . . . . . . . . . . . . . . . . . . . 10:00 am to 4:00 pm Monday-Wednesday 16-18 January . . . . . . . . . . . 7:00 am to 4:00 pm Thursday 19 January . . . . . . . . . . . . . . . . . . . . . . . . 7:00 am to Noon Speakers Audiovisual Desk Hours San Jose Convention Center, Room E Monday-Thursday 16-19 January . . . . . . . . . . . . . 7:30 am to 4:30 pm Speakers who have requested to use LCD projection from their laptop, 35mm slide projection, a VHS video player, or an overhead projector are encouraged to preview their materials at the Audiovisual Desk prior to their presentation. Speakers who have requested special equipment beyond the standard equipment noted here are asked to report to the EI '06 Audiovisual Desk upon arrival at the meeting to confirm equipment requests. Speakers will be responsible for delivering visual materials to the conference room and may retrieve their presentation materials from the room monitor in the conference room immediately following the session. Course Notes Courses will take place in various meeting rooms at the San Jose Marriott Hotel and the San Jose Convention Center. Your room assignment will be given to you with a ticket in your registration materials. Registrants for courses must exchange each course ticket received for their course notes in the course meeting room at the beginning of your class. Video/Digital Recording Policy For copyright reasons, video or digital recording of any conference session, short course, or poster session is strictly prohibited without written prior consent from each specific presenter to be recorded. Individuals not complying with this policy will be asked to leave a given session and to surrender their film or disc. It is the responsibility of the presenter to notify the conference sponsors if such consent is given.
Messages for Attendees Messages for attendees at Electronic Imaging 2006 Symposium can be left by calling the IS&T/SPIE Message Center at 408-271-6100. Messages will be taken during registration hours Sunday through Thursday. Attendees should check the message boards at the message center daily to receive their messages. Complimentary Internet Wireless Access IS&T/SPIE are pleased to provide complimentary wireless access to the Internet for all conference attendees bringing 802.11b wireless-enabled laptops or PDAs. Properly secure your computer before accessing the public wireless network. Failure to do so may allow unauthorized access to your laptop. Coverage locations and connection settings will be posted at the Registration desk. Please configure your wireless settings as follows: SSID: EI2006 (case-sensitive - all capital letters) WEP: Disabled Network Card Settings: DHCP SPIE Bookstore and Membership Booth Monday through Thursday, Open during registration hours. SPIE publishes a variety of technical books designed to meet diverse research, reference, and educational needs. Proceedings of SPIE technical conferences from this and related meetings may be purchased at the bookstore. Also available are related books in the SPIE PRESS Series, including Tutorial Texts, Milestone Series of Selected Reprints, Critical Reviews in Science & Technology, and Monographs & Handbooks. IS&T Bookstore and Membership Booth Monday through Thursday, Open during registration hours. IS&T publishes a variety of books to meet your needs. Proceedings of past IS&T conferences including Digital Printing Technologies, Color Imaging as well as Recent Progress series books will be available. IS&T also distributes selected titles from cooperating publishers of science and technology books in the imaging field. Information about upcoming IS&T conferences and IS&T membership benefits, sample journals, and newsletters are available.
8
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Cash Cart: Breakfast Breads, Snacks and Quick Lunch San Jose Convention Center, Concourse 1 Lobby Monday-Thursday 16-19 January . . . . . . . . . . . . . 7:30 am to 2:30 pm The Cash Cart will offer breakfast breads, yogurt, fruit, coffee, juice and other beverages each morning of the conference. Luncheon and snack service will include deli-style sandwiches, salads, snacks and pastries, and beverage. Attendees will need to make their own breakfast arrangements for Monday. Copy & Computer Services Marriott San Jose Hotel - Business Center. Self-service computers/ printers, fax, copiers and internet access is available on 24 hr basis. Access is by your sleeping room key. All Marriott guest rooms include T1 internet connections. San Jose Convention Center - Business Center at the administration office. Open 8:00 am to 5:00 pm daily. Services available include computers/printers, fax, and copiers. Child Care A few child sitting services available in San Jose are as follows. 1. Bay Area 2nd MOM Inc. Hotel Nanny Service Toll Free Phone: 1-888-926-3666, or (650) 858-2469, ext. 109., Monday-Friday 9:00 am to 5:00 pm. At other times phone (650) 858-4984 Fax: (650) 493-6598 E-mail: [email protected] or [email protected] Website: www.2ndmom.com 2. Sitters Unlimited Phone: (408) 452-0225 E-mail: [email protected] or www.sittersunlimited.com Note: IS&T/SPIE does not imply an endorsement or recommendation of these services. They are provided on an "information-only" basis for your further analysis and decision. Other services may be available. Car Rental Hertz Car Rental has been selected as the official car rental agency for this Symposium. To reserve a car, identify yourself as an Electronic Imaging Conference attendee using the Hertz Meeting Code CV# 029B0009. Call 1-800-654-2240.
General Information
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
9
Conference 6055A · Conv. Ctr. Room A8
Monday-Wednesday 16-18 January 2006 · Part of Proceedings of SPIE Vol. 6055 Stereoscopic Displays and Virtual Reality Systems XIII Stereoscopic Displays and Applications XVII
Conference Chairs: Andrew J. Woods, Ctr. for Marine Science and Technology/Curtin Univ. of Technology (Australia); Neil A. Dodgson, Univ. of Cambridge (United Kingdom); John O. Merritt, The Merritt Group Program Committee: Gregg E. Favalora, Actuality Systems, Inc.; Nicolas S. Holliman, Univ. of Durham (United Kingdom); Janusz Konrad, Boston Univ.; Shojiro Nagata, InterVision (Japan); Steven L. Smith, Consultant; Vivian K. Walworth, Jasper Associates; Michael A. Weissman, Micro Vision Systems Inc.
Monday 16 January SESSION 1 Conv. Ctr. Room A8 . . . . . . . . . . . . Mon. 8:30 to 10:10 am Entertainment, Visualization, and Training: Applications of Stereoscopy Chair: Andrew J. Woods, Curtin Univ. of Technology (Australia) 8:30 am: The use of stereoscopic visualization in chemistry and structural biology, M. Hus бk, Institute of Chemical Technology Prague (Czech Republic) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-01] 8:50 am: Using stereoscopic real-time graphics to shorten training time for complex mechanical tasks, F. Tecchia, M. Carrozzino, F. Rossi, M. Bergamasco, Scuola Superiore Sant'Anna (Italy); M. Vescovi, SIG Simonazzi (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-02] 9:10 am: Stereoscopic display of 3D models for design visualization, K. J. Gilson, Parsons Brinckerhoff . . . . . . . . . . . . . . . . . . . . . . . . [6055A-03] 9:30 am: Stereoscopic image production: live, CGI, and integration, E. Criado, Enxebre Entertainment (Spain) . . . . . . . . . . . . . . . . . . . [6055A-04] 9:50 am: Cosmic cookery: making a stereoscopic 3D animated movie, N. S. Holliman, C. Baugh, C. Frenk, A. Jenkins, B. Froner, D. Hassaine, J. Helly, N. Metcalfe, T. Okamoto, Univ. of Durham (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-05] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am SESSION 2 Conv. Ctr. Room A8 . . . . . . . . Mon. 10:40 am to 12:00 pm Medical Applications of Stereoscopy Chair: Michael A. Weissman, Micro Vision Systems, Inc. 10:40 am: Evaluation of stereoscopic medical video content on an autostereoscopic display for undergraduate medical education, J. F. R. Ilgner, Univ. Hospital Aachen (Germany); T. Kawai, T. Shibata, T. Yamazoe, Waseda Univ. (Japan); M. Westhofen, Univ. Hospital Aachen (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-06] 11:00 am: Stereoscopic visualization and editing of automatic abdominal aortic aneurysms (AAA) measurements for stent graft planning, L. Zhou, Y. P. Wang, C. Goh, R. Kockro, L. Serra, Volume Interactions Pte. Ltd. (Singapore) . . . . . . . . . . . . . . . . . . . . . . . [6055A-07] 11:20 am: A hybrid virtual environment for training of radiotherapy treatment of cancer., R. Phillips, The Univ. of Hull (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-08] 11:40 am: Blur spot limitations in distal endoscope sensors, A. Yaron, Visionsense Inc.; M. Shechterman, N. Horesh, U. Ronen, Visionsense Ltd. (Israel) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-09] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm
SESSION 3 Conv. Ctr. Room A8 . . . . . . . . . . . . . Mon. 1:30 to 3:10 pm Perception and Performance: Stereoscopic Human Factors Chair: John O. Merritt, The Merritt Group 1:30 pm: Visual comfort with mobile stereoscopic gaming, J. P. Hдkkinen, Nokia Research Ctr. (Finland) and Univ. of Helsinki (Finland); M. Liinasuo, Nokia Research Ctr. (Finland); J. Takatalo, G. S. Nyman, Univ. of Helsinki (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-10] 1:50 pm: Effect of disparity and motion on visual comfort of stereoscopic images, F. Speranza, J. W. Tam, R. Renaud, Communications Research Ctr. Canada (Canada); N. Hur, Electronics and Telecommunications Research Institute (South Korea) . . . . . . [6055A-11] 2:10 pm: Analysis of an autostereoscopic display: the perceptual range of the three-dimensional visual fields and saliency of static depth cues, P. R. Havig, Air Force Research Lab.; J. P. McIntire, Consortium Research Fellows Program; R. McGruder, U.S. Air Force Academy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-12] 2:30 pm: Effects of gender, application, experience, and constraints on interaction performance using autostereoscopic displays, Z. Y. Alpaslan, S. Yeh, A. A. Rizzo III, A. A. Sawchuk, Univ. of Southern California . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-13] 2:50 pm: Examination of asthenopia recovery using stereoscopic 3D display with dynamic optical correction, T. Shibata, T. Kawai, K. Ohta, L. Jae Lin, Waseda Univ. (Japan); M. Otsuki, N. Miyake, Nikon Corp. (Japan); Y. Yoshihara, Arisawa Manufacturing Co., Ltd. (Japan); T. Iwasaki, Univ. of Occupational and Environmental Health (Japan) . . . . . . . . . . . [6055A-14] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm SESSION 4 Conv. Ctr. Room A8 . . . . . . . . . . . . . . Mon. 3:40 to 5:00 pm Stereoscopic Projection and Stereoscopic Cinema Chair: Vivian K. Walworth, Jasper Associates 3:40 pm: High-resolution insets in projector-based stereoscopic displays: principles and techniques, G. Godin, P. Massicotte, L. Borgeat, National Research Council Canada (Canada) . . . . . . . . . . . . . . [6055A-15] 4:00 pm: Stereo projection using interference filters, H. Jorke, M. Fritz, Infitec GmbH (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-16] 4:20 pm: Development of the real-time stereoscopic error corrector and convergence controller, S. Nam, C. Park, Korean Broadcasting System (South Korea); Y. S. Yu, K. Lee, TVLogic Co. Ltd. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-55] 4:40 pm: 3D in digital cinema, W. J. Husak, Dolby Labs. . . . . [6055A-18]
10
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6055A · Co nv. Ctr. Room A8
3D Theatre . . . . . . . . . . . . . . . . . . . . . . 5:20 to 7:20 pm Chairs: Andrew J. Woods, Curtin Univ. of Technology (Australia); Chris Ward, Lightspeed Design, Inc. See large-screen examples of how 3D video is being used and produced around the world. SD&A Dinner . . . . . . . . . . . . . . . . . . . . 7:30 pm to Late A no-host informal dinner open to all SD&A attendees will be held at a local San Jose restaurant. Details will be available at the conference. Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 5 Conv. Ctr. Room A8 . . . . . . . . . . . . Tues. 9:30 to 10:30 am Stereoscopic Image Processing Chair: Janusz Konrad, Boston Univ. 9:30 am: Platelet-based coding of depth maps for the transmission of multiview images, Y. Morvan, D. Farin, P. H. N. de With, Technische Univ. Eindhoven (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-19] 9:50 am: Efficient view synthesis from uncalibrated stereo, R. A. C. Braspenning, M. Op de Beeck, Philips Research Labs. (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-20] 10:10 am: A fast image multiplexing method robust to viewer's position and lens misalignment in lenticular 3D displays, Y. Lee, J. B. Ra, Korea Advanced Institute of Science and Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-21] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am SESSION 6 Conv. Ctr. Room A8 . . . . . . . . Tues. 11:00 am to 12:00 pm Making Pictures: Stereoscopic Rendering Chair: Neil A. Dodgson, Univ. of Cambridge (United Kingdom) 11:00 am: Real-time rendering for multiview displays, R. M. Berretty, F. J. Peters, Philips Research Labs. (Netherlands); G. Volleberg, Philips Applied Technologies (Netherlands) . . . . . . . . . . . . . . . . . . . . . [6055A-22] 11:20 am: Anisotropic scene geometry resampling with occlusion filling for 3DTV applications, J. Kim, T. Sikora, Technische Univ. Berlin (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-23] 11:40 am: Distributed rendering for multiview parallax displays, T. Annen, Max-Planck-Institut fьr Informatik (Germany); W. Matusik, H. Pfister, Mitsubishi Electric Research Labs.; H. Seidel, Max-Planck-Institut fьr Informatik (Germany); M. Zwicker, Massachusetts Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-24] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm
SESSION 7 Conv. Ctr. Room A8 . . . . . . . . . . . . . Tues. 1:30 to 3:10 pm Autostereoscopic Displays I Chair: Gregg E. Favalora, Actuality Systems, Inc. 1:30 pm: On the number of viewing zones required for head-tracked autostereoscopic display, N. A. Dodgson, Univ. of Cambridge (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-25] 1:50 pm: Multiview LCD wall system, I. Relke, Opticality GmbH (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-26] 2:10 pm: Flatbed-type autostereoscopic display system and its image format for encoding, T. Saishu, S. Numazaki, K. Taira, R. Fukushima, A. Morishita, Y. Hirayama, Toshiba Corp. (Japan) . . . . . . . . . . . . . [6055A-27] 2:30 pm: Autostereoscopic 3D display, A. Schwerdtner, SeeReal Technologies GmbH (Germany) . . . . . . . . . . . . . . . . . . . . . . . . [6055A-28] 2:50 pm: The HoloVizio system, T. Balogh, Holografika Kft. (Hungary) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-29] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm SESSION 8 Conv. Ctr. Room A8 . . . . . . . . . . . . . Tues. 3:40 to 5:20 pm Autostereoscopic Displays II Chair: Shojiro Nagata, InterVision (Japan) 3:40 pm: Development of autostereoscopic display system for remote manipulation, T. Honda, Y. Kuboshima, K. Iwane, T. Shiina, Chiba Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-30] 4:00 pm: Ray-space acquisition and reconstruction within cylindrical objective space, T. Yendo, T. Fujii, M. Tanimoto, Nagoya Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-31] 4:20 pm: 72-directional display having VGA resolution for highappearance image generation, Y. Takaki, T. Dairiki, Tokyo Univ. of Agriculture and Technology (Japan) . . . . . . . . . . . . . . . . . . . . . [6055A-32] 4:40 pm: Combining volumetric edge display and multiview display for expression of natural 3D images, R. Yasui, I. Matsuda, H. Kakeya, Univ. of Tsukuba (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-33] 5:00 pm: Adaptive parallax control for multiview stereo panoramas, C. Wang, A. A. Sawchuk, Univ. of Southern California . . . . . . . . . [6055A-34] Posters and Demonstrations-Tuesday Chairs: Neil A. Dodgson, Univ. of Cambridge (United Kingdom); Andrew J. Woods, Curtin Univ. of Technology (Australia) Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Real-time stereographic display of voluMetric Datasets in radiology, X. H. Wang, G. S. Maitz, J. K. Leader, W. F. Good, Univ. of Pittsburgh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-46] Ergonomic evaluation system for stereoscopic video production, T. Kawai, S. Kishi, T. Yamazoe, T. Shibata, Waseda Univ. (Japan); T. Inoue, Kanagawa Institute of Technology (Japan); Y. Sakaguchi, K. Okabe, Y. Kuno, Let's Corp. (Japan); T. Kawamoto, Chukyo TV Broadcasting Corp. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-47] Wide-viewing-angle three-dimensional display system using HOE lens array, H. Takahashi, H. Fujinami, Osaka City Univ. (Japan); K. Yamada, Hiroshima Institute of Technology (Japan) . . . . . . [6055A-48]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
11
Conference 6055A · Conv. Ctr. Room A8
Depth map-based disparity estimation technique using multiview and depth camera, G. Um, Electronics and Telecommunications Research Institute (South Korea); S. Kim, K. Kim, Gwangju Institute of Science and Technology (South Korea); N. Hur, Electronics and Telecommunications Research Institute (South Korea); K. Lee, Gwangju Institute of Science and Technology (South Korea) . . . . . . . [6055A-50] A uniform metric for anaglyph calculation, Z. Zhang, D. F. McAllister, North Carolina State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-51] Multiview autostereoscopic display with double-sided reflecting scanning micromirrors, A. Nakai, K. Hoshino, K. Matsumoto, I. Shimoyama, The Univ. of Tokyo (Japan) . . . . . . . . . . . . . . . [6055A-53] Depth-enhanced floating display system based on integral imaging, J. Kim, Seoul National Univ. (South Korea); S. Min, Information and Communications Univ. (South Korea); Y. Kim, S. Cho, H. Choi, B. Lee, Seoul National Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . [6055A-54] Three-dimensional sprites for lenticular-type three-dimensional display, T. Dairiki, Y. Takaki, Tokyo Univ. of Agriculture and Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-56] Optical design considerations for a beam combiner in a StereoMirror (TM) 3D display, A. Hochbaum, VAV Consulting; J. L. Fergason, Fergason Patent Properties . . . . . . . . . . . . . . . . . [6055A-57] Horizontal parallax distortion in toe-in camera systems with fisheye lens, H. Kang, D. Kim, N. Hur, Electronics and Telecommunications Research Institute (South Korea) . . . . . . . . . . . . . . . . . . . . . [6055A-58] Implementation of 3DTV broadcasting system for realistic broadcasting services, B. Bae, S. Cho, K. Yun, H. Kang, N. Hur, C. Ahn, Electronics and Telecommunications Research Institute (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-59] Performance analysis of a compact electro-optical 3D adapter with a wide capturing angle, S. Kim, J. Lee, E. Kim, Kwangwoon Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-61] New method of zoom-convergence interlocked control in the moving parallel-axes style stereoscopic camera, J. Lee, S. Nam, J. Lee, C. Park, S. Chung, Korean Broadcasting System (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-62]
Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 9 Conv. Ctr. Room A8 . . . . . . . . . . . . Wed. 9:30 to 10:30 am Integral 3D Imaging Chair: Shojiro Nagata, InterVision (Japan) 9:30 am: Integral videography of high-density light field with spherical layout camera array, T. Koike, M. Oikawa, N. Kimura, F. Beniyama, T. Moriya, M. Yamasaki, Hitachi, Ltd. (Japan) . . . . . . . . . . . . . . . [6055A-35] 9:50 am: Imaging properties of microlens arrays for integral imaging system, J. Arai, M. Okui, Y. Nojiri, F. Okano, NHK Science & Technical Research Labs. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-36] 10:10 am: Comparative study on 3D-2D convertible integral imaging systems, B. Lee, H. Choi, J. Kim, Y. Kim, Seoul National Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-37] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am Discussion Forum: Stereoscopic Digital Cinema: The Way of the Future or a 9-Day Wonder? Moderator: Charles Swartz, Entertainment Technology Ctr./Univ. of Southern California Panel Members: Lenny Lipton, CTO, REAL D; Ray Zone, The 3-D Zone; John Rupkalvis, StereoScope International; Walter Husak, Dolby Labs Inc.; Neil Feldman, In-Three, Inc. Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:50 pm SESSION 10 Conv. Ctr. Room A8 . . . . . . . . . . . . . Wed. 1:50 to 2:10 pm Stereoscopic Software Chair: Nicolas S. Holliman, Univ. of Durham (United Kingdom) 1:50 pm: Application of 3DHiVision: a system with a new 3D HD renderer, P. Sun, Sun Advanced Engineering, Inc. (Japan); S. Nagata, InterVision (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-39]
12
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6055A · Conv. Ctr. Room A8 SESSION 11 Conv. Ctr. Room A8 . . . . . . . . . . . . . Wed. 2:20 to 3:20 pm Keynote Presentation Chair: Andrew J. Woods, Curtin Univ. of Technology (Australia) Keynote 3D animation in three dimensions: the rocky road to the obvious, Hugh Murray, IMAX Corp. (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-40] That animation created using CG modeling and animation tools is inherently three-dimensional is well known. In the middle to late nineties IMAX Corporation began actively exploring CG animated features as a possible source of economically viable content for its rapidly growing network of stereoscopic IMAX(r) 3D theatres. The journey from there to the spectacular success of the IMAX(r) 3D version of The Polar Express is an interesting mix of technical, creative and production challenges. For example 3D animations often have 2D elements and include many sequences that have framing, composition and lens choices that a stereographer would have avoided had 3D been part of the recipe at the outset. And of course the decision to ask for a second set of deliverables from an already stressed production takes nerve. The talk will cover several of these issues and explain why the unique viewing experience enabled by the wide-angle geometry of IMAX(r) 3D theatres makes it worth all the pain. Biography: Hugh Murray is Vice President, Technical Production at IMAX Corporation and for the past 10 years he has worked on most of IMAX Corporation's film productions as a technical expert, particularly in 3D and special effects. Hugh was the instigator and producer, with Steve Hoban, of the computer animated film Cyberworld 3D. Hugh led the team that identified the key technologies for the IMAX DMR(r) process and was IMAX(r) Producer on Apollo 13 (with Lorne Orleans), Star Wars Episode II: Attack of the Clones (with Lorne Orleans), Matrix Reloaded, Matrix Revolutions, and the IMAX(r) 3D version of The Polar Express. He was most recently an Executive Producer on Magnificent Desolation: Walking On The Moon 3D and is currently working on the 3D versions of two animated features for release in 2006. Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:20 to 3:50 pm SESSION 12 Conv. Ctr. Room A8 . . . . . . . . . . . . . Wed. 3:50 to 5:30 pm Stereoscopic Developments Chair: Steven L. Smith, Consultant 3:50 pm: A method of real-time construction of full parallax light field, K. Tanaka, S. Aoki, Sony Corp. (Japan) . . . . . . . . . . . . . . . . . . [6055A-41] 4:10 pm: Depth maps created from blur information using images with focus at near and at far, S. Cho, Electronics and Telecommunications Research Institute (South Korea); J. W. Tam, F. Speranza, R. Renaud, Communications Research Ctr. Canada (Canada); N. Hur, S. Lee, Electronics and Telecommunications Research Institute (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-49] 4:30 pm: Simulation of 3D image depth perception in a 3D display using two stereoscopic displays at different depths, K. Uehira, Kanagawa Institute of Technology (Japan) . . . . . . . . . . . . . . . . [6055A-43] 4:50 pm: Innovative stereoscopic display using variable polarized angle, J. E. Gaudreau, PolarScreens, Inc. (Canada) and MacNaughton, Inc.; M. Bechamp, PolarScreens, Inc. (Canada); B. MacNaughton, V. S. Power, MacNaughton, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055A-44] 5:10 pm: A novel walk-through 3D display, S. DiVerdi, A. Olwal, I. K. Rakkolainen, T. Hцllerer, Univ. of California/Santa Barbara . . . [6055A-45]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
13
Conference 6055B · Conv. Ctr. Room B3
Thursday 19 January 2006 · Part of Proceedings of SPIE Vol. 6055 Stereoscopic Displays and Virtual Reality Systems XIII The Engineering Reality of Virtual Reality 2006 Conference Chairs: Mark T. Bolas, Univ. of Southern California; Ian E. McDowall, Fakespace Labs., Inc. Program Committee: Nick England, 3rdTech, Inc.; Guillaume Moreau, CNRS Aeronmie (France); Shojiro Nagata, InterVision (Japan); Daniel J. Sandin, Univ. of Illinois/Chicago; Andreas Simon, Fraunhofer Institute for Media Communication (Germany); Henry A. Sowizral, Sun Microsystems, Inc.
Thursday 19 January SESSION 13 Conv. Ctr. Room B3 . . . . . . . . . . . . Thurs. 8:30 to 9:50 am Procedures Chair: Ian E. McDowall, Fakespace Labs., Inc. 8:30 am: Texturing of continuous LoD meshes with the hierarchical texture atlas, H. Birkholz, Univ. Rostock (Germany) . . . . . . . . [6055B-63] 8:50 am: Optimal approximation of head-related transfer function's zero-pole model based on genetic algorithm, J. Zhang, Southeast Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055B-64] 9:10 am: Multiprojector image distortion correction scheme for curved screens on the example of the Cybersphere, B. V. Shulgin, J. Ye, V. H. Raja, Univ. of Warwick (United Kingdom) . . . . . . . . . . . . . . . . . [6055B-65] 9:30 am: 3D workflow for HDR image capture of projection systems and objects for CAVE virtual environments authoring with wireless touch-sensitive devices, M. J. Prusten, Optical Design Labs.; M. K. McIntyre, Total Eclipse Studios; M. Landis, The Univ. of Arizona [6055B-66] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9:50 to 10:20 am SESSION 14 Conv. Ctr. Room B3 . . . . . . . . . . Thurs. 10:20 to 11:40 am Applications Chair: Ian E. McDowall, Fakespace Labs., Inc. 10:20 am: Examination of corner vane estrangement evaluation method for a circular tunnel, H. Yokoyama, O. Fujishima, Hitachi, Ltd. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055B-67] 10:40 am: Virtual technical support for field engineers in the water and ventilation hygiene industry, I. A. Nicholas, Cardiff Univ. (United Kingdom); D. Kim, Aqua Marc Ltd. (United Kingdom) . . . . . . . [6055B-68] 11:00 am: Virtual reality in construction industry: a requirement compatibility analysis approach, J. Ye, B. V. Shulgin, V. H. Raja, Univ. of Warwick (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055B-69] 11:20 am: Adding tactile realism to a virtual reality laparoscopic surgical simulator with a cost-effective human interface device, I. W. Mack, Queen's Univ. Belfast (United Kingdom); S. Potts, The Royal Group of Hospitals (United Kingdom); K. R. McMenemy, R. S. Ferguson, Queen's Univ. Belfast (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . [6055B-70] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:40 am to 1:10 pm
SESSION 15 Conv. Ctr. Room B3 . . . . . . . . . . . . Thurs. 1:10 to 2:30 pm The Medium Chair: Mark T. Bolas, Fakespace Labs., Inc. 1:10 pm: Inverse perspective, M. Dolinsky, Indiana Univ. . . . [6055B-71] 1:30 pm: Virtual reality and the unfolding of higher dimensions, J. C. Aguilera, Univ. of Illinois at Chicago . . . . . . . . . . . . . . . . . . . . . [6055B-72] 1:50 pm: Framing the magic, D. Tsoupikova, Univ. of Illinois at Chicago . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055B-73] 2:10 pm: Virtual reality, immersion, and the unforgettable experience, J. F. Morie, Univ. of Southern California . . . . . . . . . . . . . . . . . . [6055B-74] SESSION 16 Conv. Ctr. Room B3 . . . . . . . . . . . . Thurs. 2:30 to 4:40 pm Viewpoints Chair: Mark T. Bolas, Fakespace Labs., Inc. 2:30 pm: Teleoperation interface for mobile robot with perspectivetransformed virtual 3D screen on PC display, T. Kimura, H. Kakeya, Univ. of Tsukuba (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6055B-75] 2:50 pm: An orientation sensing interface for portable situation awareness displays, J. Bleecker, Univ. of Southern California [6055B-76] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm 3:40 pm: An interactive camera placement and visibility simulator for image-based VR applications, A. State, G. Welch, A. Ilie, The Univ. of North Carolina at Chapel Hill . . . . . . . . . . . . . . . . . . . . . . . . . . [6055B-77] 4:00 pm: Overview of virtual camera mechanisms for collaborative virtual environments: an application to the VRIMOR project, E. E. Alvarez, A. A. De Antonio, Univ. Politйcnica de Madrid (Spain) [6055B-78] 4:20 pm: Synthecology: `sound' use of audio in teleimmersion, G. A. Baum, SUNY/Univ. at Buffalo; M. Gotsis, Univ. of Southern California; B. Chang, R. Drinkwater, D. St. Clair, Art Institute of Chicago . . . [6055B-79] Panel Discussion . . . . . . . . . . . . . . . . . . . . . 4:40 to 5:40 pm
14
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6056 · Conv. Ctr. Room C1
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6056 Three-Dimensional Image Capture and Applications VI
Conference Chairs: Brian D. Corner, U.S. Army Natick Soldier Ctr.; Peng Li, GEO-Centers, Inc.; Matthew Tocheri, Arizona State Univ.
Monday 16 January SESSION 1 Conv. Ctr. Room C1 . . . . . . . . . . . . Mon. 8:40 to 10:00 am 3D Scanning Hardware Chair: Brian D. Corner, U.S. Army Natick Soldier Ctr. 8:40 am: A novel design of grating projecting system for 3D reconstruction of wafer bumps, Y. Shu, Xi'an Jiaotong Univ. (China); R. C. Chung, J. Cheng, The Chinese Univ. of Hong Kong (Hong Kong China); E. Y. Lam, The Univ. of Hong Kong (Hong Kong China); K. S. M. Fung, F. Wang, ASM Assembly Automation Ltd. (Hong Kong China) . . . . [6056-01] 9:00 am: Measurement of discontinuities on 3D objects using digital moirй, J. Liao, L. Cai, The Hong Kong Univ. of Science and Technology (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-02] 9:20 am: High-speed and high-sensitive demodulation pixel for 3D imaging, B. Bьttgen, T. Oggier, Ctr. Suisse d'Electronique et de Microtechnique SA (Switzerland); P. Seitz, Swissnex; F. Lustenberger, Ctr. Suisse d'Electronique et de Microtechnique SA (Switzerland) . . . [6056-03] 9:40 am: A QVGA-size CMOS time-of-flight range image sensor with background light charge draining structure, T. Ushinaga, I. Abdul Halin, T. Sawada, S. Kawahito, Shizuoka Univ. (Japan); M. Homma, Sharp Corp. (Japan); Y. Maeda, Suzuki Motor Corp. (Japan) . . . . . . . . . . . . . . [6056-04] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am SESSION 2 Conv. Ctr. Room C1 . . . . . . . . Mon. 10:30 am to 12:00 pm 3D Object Capture from Static Scans and Video I Chair: Brian D. Corner, U.S. Army Natick Soldier Ctr. 10:30 am: Overview of 3D surface digitization technologies in Europe (Invited Paper), N. D'Apuzzo, Homometrica Consulting (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-05] 11:00 am: Automatic 3D real world surface texture mapping using perspective projection methods, C. Shih, MingDao Univ. (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-06] 11:20 am: Virtual confocal macroscopy, P. M. Hanna, U.S. Air Force; B. D. Rigling, Wright State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-07] 11:40 am: A robust algorithm for estimation of depth map for 3D shape recovery, A. Malik, T. Choi, Gwangju Institute of Science and Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-08] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm
SESSION 3 Conv. Ctr. Room C1 . . . . . . . . . . . . . Mon. 1:30 to 5:00 pm 3D Object Capture from Static Scans and Video II Chair: Peng Li, GEO-Centers, Inc. 1:30 pm: Formation of stereoscopic image pairs from a sequence of frames, M. A. Wessels, Dimensional Imaging, LLC . . . . . . . . . . . [6056-09] 1:50 pm: 3D model generation using unconstrained motion of a handheld video camera, C. Baker, C. H. Debrunner, M. Whitehorn, PercepTek, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-10] 2:10 pm: 3D from arbitrary 2D video, I. A. Ideses, L. P. Yaroslavsky, TelAviv Univ. (Israel) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-11] 2:30 pm: Nonintrusive viewpoint tracking for 3D for perception in smart video conference, X. Desurmont, I. Ponte, J. Meessen, J. Delaigle, Multitel A.S.B.L. (Belgium) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-12] 2:50 pm: Internal shape-deformation invariant 3D surface matching using 2D principal component analysis, M. Celenk, I. Al-Jarrah, Ohio Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-13] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm 3:40 pm: Digital Hammurabi: design and development of a 3D scanner for cuneiform tablets, D. V. Hahn, D. Duncan, K. Baldwin, J. Cohen, B. Purnomo, Johns Hopkins Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-14] 4:00 pm: Three-dimensional surface reconstruction for evaluation of the abrasion effects on textile fabrics, A. O. Mendes, P. T. Fiadeiro, R. A. Miguel, Univ. da Beira Interior (Portugal) . . . . . . . . . . . . . . . . . . . [6056-15] 4:20 pm: 3D environment capture from monocular video and inertial data, R. Clark, M. Lin, C. J. Taylor, Acuity Technology . . . . . . . . . [6056-16] 4:40 pm: The effects of different shape-based metrics on identification of military targets from 3D ladar data, G. J. Meyer, J. Weber, Air Force Research Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-18] Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 4 Conv. Ctr. Room C1 . . . . . . . . . Tues. 9:30 am to 12:10 pm 3D Scans of the Human I Chair: Matthew Tocheri, Arizona State Univ. 9:30 am: Digital 3D reconstruction of George Washington (Invited Paper), A. Razdan, Arizona State Univ. . . . . . . . . . . . . . . . . . . . . [6056-19] 10:00 am: The study of craniofacial growth patterns using 3D laser scanning and geometric morphometrics, M. Friess, Anthrotech Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-20] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am
electronicimaging.org · Tel: +1 703 642 9090 · [email protected].org
15
Conference 6056 · Conv. Ctr. Room C1
10:50 am: A three-dimensional analysis of the geometry and curvature of the proximal tibial articular surface of hominoids, E. K. Landis, P. A. Karnick, Arizona State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-21] 11:10 am: New approach in curve matching technique and its implications on human evolution research, H. Vahdati, P. A. Karnick, Arizona State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-22] 11:30 am: Point cloud-based 3D head model classification using optimized EGI, X. Tong, H. Wong, B. Ma, City Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-23] 11:50 am: 3D face structure extraction using shape matching morphing model, F. Xue, X. Ding, Tsinghua Univ. (China) . . . . . . [6056-24] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:40 pm SESSION 5 Conv. Ctr. Room C1 . . . . . . . . . . . . . Tues. 1:40 to 2:40 pm 3D Scans of the Human II Chair: Brian D. Corner, U.S. Army Natick Soldier Ctr. 1:40 pm: Posture and re-positioning considerations of a torso imaging system for assessing scoliosis, P. O. Ajemba, N. G. Durdle, Univ. of Alberta (Canada); D. L. Hill, J. V. Raso, Glenrose Rehabilitation Hospital (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-25] 2:00 pm: Reverse engineering and rapid prototyping techniques to innovate prosthesis socket design, G. Magrassi, G. Colombo, M. Bertetti, D. Bonacini, Politecnico di Milano (Italy) . . . . . . . . . . . . . [6056-26] 2:20 pm: Measuring human movement for biomechanical applications using markerless motion capture, L. Mьndermann, S. Corazza, A. M. Chaudhari, T. P. Andriacchi, Stanford Univ.; A. Sundaresan, R. Chellappa, Univ. of Maryland/College Park . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-28] Posters and Demonstrations-Tuesday Chair: Peng Li, GEO-Centers, Inc. Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Development of measurement system of three-dimensional shape and surface reflectance, T. Miyasaka, K. Araki, Chukyo Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-29] Use of laser 3D digitizer in data collection and 3D modeling of anatomical structures, K. Tse, Univ. of New South Wales (Australia); H. Van Der Wall, Concord Repatriation General Hospital (Australia); D. H. Vu, Univ. of New South Wales (Australia) . . . . . . . . . . . . . . . . . [6056-30] Volume intersection with imprecise camera parameters, S. Sakamoto, K. Shoji, H. Iwase, F. Toyama, J. Miyamichi, Utsunomiya Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-31] Development of ultrathin three-dimensional image capturing system, K. Yamada, H. Mitsui, T. Asano, Hiroshima Institute of Technology (Japan); H. Takahashi, Osaka City Univ. (Japan); J. Tanida, Osaka Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-32]
Run-based volume intersection for shape recovery of objects from their silhouettes, K. Shoji, S. Sakamoto, H. Iwase, F. Toyama, J. Miyamichi, Utsunomiya Univ. (Japan) . . . . . . . . . . . . . . . . . . . . [6056-33] A prototype system for 3D measurement using flexible calibration method, M. Fukuda, T. Miyasaka, K. Araki, Chukyo Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-34] Estimation of object motion with known structure in moving camera, H. J. Kwon, N. Hur, S. Lee, Electronics and Telecommunications Research Institute (South Korea) . . . . . . [6056-35] Synthesizing wide-angle and arbitrary view-point images from a circular camera array, N. Fukushima, T. Yendo, T. Fujii, M. Tanimoto, Nagoya Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-36] 3D urban scene reconstruction from high-resolution IKONOS stereo images, M. Fedi, T. Riadh, B. Ziad, SUPCOM (Tunisia) . . . . . . [6056-37] Procedure and algorithm of 3D reconstruction of large-scale ancient architecture, S. Xia, Y. Zhu, X. Li, Wuhan Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6056-38] Real-time 3D image-guided patient positioning in radiation therapy, D. Liu, Henry Ford Health System; G. Yin, Genex Technologies, Inc. (China); S. Li, Henry Ford Health System . . . . . . . . . . . . . . . . . [6056-41]
16
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6057 · Conv. Ctr. Room A3
Monday-Thursday 16-19 January 2006 · Proceedings of SPIE Vol. 6057 Human Vision and Electronic Imaging XI
Conference Chairs: Bernice E. Rogowitz, IBM Thomas J. Watson Research Ctr.; Thrasyvoulos N. Pappas, Northwestern Univ.; Scott J. Daly, Sharp Labs. of America, Inc. Program Committee: Albert J. Ahumada, Jr., NASA Ames Research Ctr.; Jan P. Allebach, Purdue Univ.; Walter R. Bender, MIT Media Lab.; Michael H. Brill, Datacolor; John C. Dalton, Synthetik Software; Huib de Ridder, Technische Univ. Delft (Netherlands); Gunilla A. M. Derefeldt, Swedish Defence Research Agency (Sweden); Miguel P. Eckstein, Univ. of California/Santa Barbara; Elena A. Fedorovskaya, Eastman Kodak Co.; Jennifer Gille, Raytheon Co.; Laurent Itti, Univ. of Southern California; Stanley A. Klein, Univ. of California/Berkeley; Jan J. Koenderink, Univ Utrecht (Netherlands); John J. McCann, McCann Imaging; Jeffrey B. Mulligan, NASA Ames Research Ctr.; Karol Myszkowski, Max-Planck-Institut fьr Informatik (Germany); Adar Pelah, The Univ. of York (United Kingdom); Hawley K. Rising III, Sony Electronics; Robert J. Safranek, Benevue, Inc.; Christopher W. Tyler, Smith-Kettlewell Institute; Andrew B. Watson, NASA Ames Research Ctr. Note: Please see room sign for papers added after program was printed.
Monday 16 January SESSION 1 Conv. Ctr. Room A3 . . . . . . . . . Mon. 9:00 am to 12:00 pm Keynote Session 10:00 am: Computational Neuroimaging: maps and Tracks in the Human Brain, B. Wandell, Stanford University 11:00 am: Learning where to look, Mary Hayhoe, University of Rochester Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm SESSION 2 Conv. Ctr. Room A3 . . . . . . . . . . . . . Mon. 1:30 to 3:30 pm Mechanisms of Luminance, Color, and Temporal Sensitivity 1:30 pm: Local luminance effect on spatial summation in the foveal vision and its implication on image artifact classification, C. Chen, S. Y. Lin, H. G. Han, National Taiwan Univ. (Taiwan) . . . . . . . . . . . . . . . [6057-01] 1:50 pm: Evaluating contrast sensitivity, S. Kitagura, L. W. MacDonald, London College of Communication (United Kingdom) . . . . . . . . . [6057-02] 2:10 pm: Spatio-velocity CSF as a function of retinal velocity using unstabilized stimuli, J. L. Laird, M. R. Rosen, J. B. Pelz, E. D. Montag, Rochester Institute of Technology; S. J. Daly, Sharp Labs. of America, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-03] 2:30 pm: A basis for cones, B. V. Funt, W. Xiong, Simon Fraser Univ. (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-04] 2:50 pm: High-dynamic range scene compression in humans, J. J. McCann, McCann Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-47] 3:10 pm: Computational model of lightness perception in HDR imaging, G. Krawczyk, K. Myszkowski, H. Seidel, Max-Planck-Institut fьr Informatik (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-05] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:30 to 4:00 pm
SESSION 3 Conv. Ctr. Room A3 . . . . . . . . . . . . . Mon. 4:00 to 6:00 pm Eye Movements, Visual Search, and Attention: A Tribute to Larry Stark 4:00 pm: Larry Stark and scan path, S. R. Ellis, NASA Ames Research Ctr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-46] 4:20 pm: A new metrics for definition of gaze area from the geometrical structures of picture composition, M. Yamazaki, M. Kameda, Iwate Prefectural Univ. (Japan) . . . . . . . . . . . . . . . . . . . [6057-06] 4:40 pm: Target salience and visual search on novel and familiar backgrounds, K. McDermott, Univ. of Nevada/Reno; J. Mulligan, NASA Ames Research Ctr.; G. Bebis, M. Webster, Univ. of Nevada/ Reno . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-43] HVEI Banquet . . . . . . . . . . . . . . . . . . . . 7:30 to 10:30 pm The annual Human Vision and Electronic Imaging Banquet will be held Monday evening, 16 January, 7:30 to 10:30 pm. The banquet will take place in a local restaurant or wine cellar. For tickets and more information, please visit the Electronic Imaging Registration Desk. Banquet Speaker: Jack Tumblin, Northwestern Univ. Rethinking Photography: Digital Devices to Capture Appearance
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
17
Conference 6057 · Conv. Ctr. Room A3
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 4 Conv. Ctr. Room A3 . . . . . . . . . Tues. 9:30 am to 12:00 pm Perceptual Image Quality and Applications 9:30 am: Effects of spatial correlations and global precedence on the visual fidelity of distorted images, D. M. Chandler, K. H. S. Lim, S. S. Hemami, Cornell Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-07] 9:50 am: Pseudo no reference image quality metric using perceptual data hiding, A. Ninassi, P. Le Callet, F. Autrusseau, Univ. de Nantes (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-08] 10:10 am: Attention-based color correction, F. W. M. Stentiford, Univ. College London (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . [6057-09] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am 11:00 am: Contrast enhancement of medical images using multiscale decomposition, M. A. Trifas, J. M. Tyler, O. S. Pianykh, Louisiana State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-10] 11:20 am: Human visual alpha stable models for digital halftoning, A. J. Gonzбlez, J. Bacca Rodrнguez, G. R. Arce, Univ. of Delaware; D. L. Lau, Univ. of Kentucky . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-11] 11:40 am: Study of asthenopia caused by the viewing of stereoscopic images, H. Hagura, Tokyo Institute of Technology (Japan) . . . . . [6057-12] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm SESSION 5 Conv. Ctr. Room A3 . . . . . . . . . . . . . Tues. 1:30 to 2:50 pm Visually Tuned Algorithms for the Design and Analysis of Flat-Panel Displays 1:30 pm: Perceptual image quality improvement for large screen displays, F. Lebowsky, Y. Huang, H. Wang, STMicroelectronics (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-13] 1:50 pm: LCD motion-blur analysis, perception, and reduction using synchronized backlight flashing, X. Feng, Sharp Labs. of America, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-14] 2:10 pm: Human vision-based algorithm to hide defective pixels in LCDs, T. R. Kimpe, S. Coulier, Barco N.V. (Belgium) . . . . . . . . . . [6057-15] 2:30 pm: Using optimal rendering to visually mask defective subpixels, D. S. Messing, L. J. Kerofsky, Sharp Labs. of America, Inc. . . . . [6057-16] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2:50 to 3:20 pm
SESSION 6 Conv. Ctr. Room A3 . . . . . . . . . . . . . Tues. 3:20 to 5:40 pm Perceptual Issues in Video Quality 3:20 pm: Perceptual study of the impact of varying frame rate on motion imagery interpretability quality, C. P. Fenimore, National Institute of Standards and Technology; J. M. Irvine, Science Applications International Corp.; D. Cannon, National Geospatial-Intelligence Agency; J. W. Roberts, I. Aviles, National Institute of Standards and Technology; S. A. Israel, Science Applications International Corp.; M. Brennan, The Boeing Co.; L. Simon, J. R. Miller, D. S. Haverkamp, Science Applications International Corp.; P. F. Tighe, M. Gross, Booz Allen Hamilton Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-17] 3:40 pm: Color preference and perceived color naturalness of digital videos, C. C. Koh, J. M. Foley, S. K. Mitra, Univ. of California/Santa Barbara . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-18] 4:00 pm: Stabilizing viewing distances in subjective assessments of mobile video, M. D. Brotherton, British Telecommunications plc (United Kingdom); K. Brunnstrцm, Acreo AB (Sweden); D. Hands, British Telecommunications plc (United Kingdom) . . . . . . . . . . . . . . . . . [6057-19] 4:20 pm: Predicting subjective video quality from separated spatial and temporal assessment, R. R. Pastrana-Vidal, J. Gicquel, J. Blin, France Telecom R&D (France); H. Cherifi, Univ. de Bourgogne (France) [6057-20] 4:40 pm: Handling of annoying variations of performances in video algorithm optimization, M. M. Nicolas, STMicroelectronics (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-21] 5:00 pm: Structural similarity quality metrics in a coding context: exploring the space of realistic distortions, A. Brooks, T. N. Pappas, Northwestern Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-22] 5:20 pm: Lossy compression of high dynamic range images and video, R. Mantiuk, K. Myszkowski, H. Seidel, Max-Planck-Institut fьr Informatik (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-23] Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Psychophysical measurement for perceptual image brightness enhancement based on image classification, I. Kim, W. Choe, S. Lee, SAMSUNG Advanced Institute of Technology (South Korea) . [6057-34] Visual deficiency and image recognition: an image semantic cartography related to visual performance, A. Scherlen, J. Da Rugna, Univ. Jean Monnet Saint-Etienne (France) . . . . . . . . . . . . . . . . [6057-35] Simple color conversion method to perceptible images for color vision deficiencies, M. Meguro, C. Takahashi, T. Koga, Yamagata Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-36] Toward a taxonomy of textures for image retrieval, J. S. Payne, Buckinghamshire Chilterns Univ. College (United Kingdom); T. J. Stonham, Brunel Univ. (United Kingdom) . . . . . . . . . . . . . . . . . [6057-38]
18
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6057 · Conv. Ctr. Room A3
Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity, M. Phielipp, J. A. Black, Jr., S. Panchanathan, Arizona State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-39] Subjective video quality evaluation for multimedia applications, Q. Huynh-Thu, Psytechnics Ltd. (United Kingdom); M. Ghanbari, Univ. of Essex (United Kingdom); D. Hands, M. D. Brotherton, British Telecommunications plc (United Kingdom) . . . . . . . . . . . . . . . [6057-40] Texture segmentation using adaptive Gabor filters based on HVS, S. Bi, D. Liang, Dalian Maritime Univ. (China) . . . . . . . . . . . . . . . . [6057-41] Image stability analysis on the human retina, M. I. Baritz, Univ. Transilvania din Brasov (Romania) . . . . . . . . . . . . . . . . . . . . . . [6057-42] Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 7 Conv. Ctr. Room A3 . . . . . . . . . Wed. 9:30 am to 12:00 pm Perceptual Approaches to Image Analysis 9:30 am: A closer look at texture metrics, H. H. Shenas, V. Interrante, Univ. of Minnesota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-24] 9:50 am: M-HinTS: mimicking humans in texture sorting, E. L. van den Broek, Vrije Univ. Amsterdam (Netherlands); T. Kok, T. E. Schouten, Radboud Univ. Nijmegen (Netherlands); E. M. van Rikxoort, Univ. Medisch Ctr. Utrecht (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-25] 10:10 am: Inference and segmentation in cortical processing, Y. Liu, G. A. Cecchi, A. R. Rao, J. Kozloski, C. Peck, IBM Thomas J. Watson Research Ctr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-26] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am 11:00 am: Subjective segmentation evaluation methods: a survey, E. Drelie Gelasca, T. Ebrahimi, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-27] 11:20 am: Perceptually based techniques for semantic image classification and retrieval, D. Depalov, T. N. Pappas, Northwestern Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-28] 11:40 am: Is Wцlfflinґs system for characterizing art possible to validate by methods used in cognitive-based image-retrieval (CBIR)?, G. A. M. Derefeldt, S. Nyberg, J. Alfredson, H. U. Allberg, Swedish Defence Research Agency (Sweden) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-29] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:40 pm
SESSION 8 Conv. Ctr. Room A3 . . . . . . . . . . . . . Wed. 1:40 to 2:40 pm Detection, Recognition, and Navigation in Complex Environments 1:40 pm: Symbol discriminability models for improved flight displays, A. J. Ahumada, Jr., M. Trujillo San-Martin, J. Gille, NASA Ames Research Ctr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-30] 2:00 pm: Is haptic watermarking worth it?, M. Barni, D. Prattichizzo, G. Menegaz, A. Formaglio, M. Franzini, Univ. degli Studi di Siena (Italy); H. Z. Tan, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-32] 2:20 pm: Display conditions that influence wayfinding in virtual environments, R. A. Browse, D. W. S. Gray, Queen's Univ. (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6057-33] Natural Image Statistics Conv. Ctr. Room A3 . . . . . . . . . . . . . Wed. 2:40 to 4:30 pm
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
19
Conference 6058 · Conv. Ctr. Room A2
Tuesday-Thursday 17-19 January 2006 · Proceedings of SPIE Vol. 6058 Color Imaging XI: Processing, Hardcopy, and Applications
Conference Chairs: Reiner Eschbach, Xerox Corp.; Gabriel G. Marcu, Apple Computer, Inc. Program Committee: A. Ufuk Agar, Garanti Technology (Turkey); Jan P. Allebach, Purdue Univ.; Jan Bares, NexPress Solutions, LLC; Phil J. Green, London College of Communication (United Kingdom); Roger David Hersch, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland); Patrick G. Herzog, GretagMachbeth (Germany); Choon-Woo Kim, Inha Univ. (South Korea); Michael A. Kriss, Consultant; Shaun T. Love, Lexmark International, Inc.; Alessandro Rizzi, Univ. Degli Studi di Milano (Italy); Shoji Tominaga, Osaka ElectroCommunication Univ. (Japan); Chris Tuijn, Agfa-Gevaert NV (Belgium)
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 1 Conv. Ctr. Room A2 . . . . . . . . . . . . . Tues. 3:30 to 6:00 pm Color and Color Transforms Chair: Reiner Eschbach, Xerox Corp. 3:30 pm: Ideal illuminants for rod/L-cone color, J. J. McCann, McCann Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-01] 4:00 pm: Accuracy of color transforms, P. J. Green, London College of Communication (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . [6058-02] 4:20 pm: Color image dequantization by constrained diffusion, D. Keysers, T. M. Breuel, DFKI GmbH (Germany) and Univ. of Kaiserslautern (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-03] 4:40 pm: Spring-primary mapping: a fast color mapping method for primary adjustment and gamut mapping, H. Zeng, Hewlett-Packard Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-04] 5:00 pm: A framework for image-dependent gamut mapping, J. Giesen, E. Schuberth, ETH Zьrich (Switzerland); K. Simon, EMPA (Switzerland); D. Zeiter, ETH Zьrich (Switzerland); P. Zolliker, EMPA (Switzerland) . [6058-05] 5:20 pm: Perceptual gamut mapping algorithm development based upon image quality and preference factors, B. Kang, M. Cho, H. Choh, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-06] 5:40 pm: Gamut estimation using 2D surface splines, M. Q. Shaw, Hewlett-Packard Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-07]
Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. High-resolution image viewing on multiprojector distributed parallel rendering display wall, J. Meng, H. Lin, J. Shi, Zhejiang Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-40] The application of wavelet transforms and mathematics morphology on the processing of infrared satellite cloud image, J. Xue, Z. Liu, P. Wang, Tianjin Univ. (China) . . . . . . . . . . . . . . . . [6058-41] Simulation and parameter optimizing of multielectrode capacitive transducers based on finite element method, D. Chen, Harbin Univ. of Science and Technology (China) . . . . . . . . . . . . . . . . . . . . . [6058-42] Digital watermarking of color image, S. C. Chao, Ta Hwa Institute of Technology (Taiwan); H. M. Huang, Tung Nan Institute of Technology (Taiwan); C. Y. Chen, Institute of Nuclear Energy Research (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-43] Subjective assessment of printed color image quality as saturation of the primary colors RGB decreases, W. Song, H. Seki, Naltec Inc. (Japan); G. Ohashi, Y. Shimodaira, Shizuoka Univ. (Japan) . . . [6058-45] Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 2 Conv. Ctr. Room A2 . . . . . . . . . . . . Wed. 9:30 to 10:20 am Color Adjustment Chair: Phil J. Green, London College of Communication (United Kingdom) 9:30 am: Uncalibrated color (Invited Paper), N. Moroney, HewlettPackard Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-08] 10:00 am: Geometrical methods for lightness adjustment in YCC color spaces, R. Samadani, G. Li, Hewlett-Packard Labs. . . . . . . . . . . [6058-09] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am
20
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6058 · Conv. Ctr. Room A2
SESSION 3 Conv. Ctr. Room A2 . . . . . . . . Wed. 10:50 am to 12:10 pm Digital Photography and Applications Chair: Michael A. Kriss, Consultant 10:50 am: Measuring gloss by digital photography, P. Kumar, L. W. MacDonald, London College of Communication (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-10] 11:10 am: Ubiquitous image processing: a novel image-enhancement facility for consumers, R. Shaw, P. Johnson, White Rose Digital [6058-11] 11:30 am: Color constancy on Japanese animation, Y. G. Ichihara, Hosen-Gakuen College (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-12] 11:50 am: Convincing non-printers to become future customers, R. Fageth, W. Schmidt-Sacht, CeWe Color AG & Co. OHG (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-13] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:30 pm SESSION 4 Conv. Ctr. Room A2 . . . . . . . . . . . . . Wed. 1:30 to 3:00 pm Displays I Chair: Choon-Woo Kim, Inha Univ. (South Korea) 1:30 pm: DTV color and image processing: past, present, and future (Invited Paper), C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-14] 2:00 pm: Subpixel rendering method for color error minimization on subpixel structured display, W. Choe, S. Lee, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) . . . . . . . . . . . . . [6058-15] 2:20 pm: Compensation method for color defects in PDP due to different time responses of phosphors, H. Oh, H. Lee, D. Park, S. Kim, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-16] 2:40 pm: Six-primary-color LCD monitor using six-color LEDs with an accurate calibration system, H. Sugiura, H. Kaneko, S. Kagawa, J. Someya, H. Tanizoe, Mitsubishi Electric Corp. (Japan) . . . . . . . . [6058-17] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm SESSION 5 Conv. Ctr. Room A2 . . . . . . . . . . . . . Wed. 3:30 to 5:30 pm Displays II Chair: Gabriel G. Marcu, Apple Computer, Inc. 3:30 pm: A color control method for image output with projection displays, S. Tominaga, K. Kumamoto, Osaka Electro-Communication Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-18] 3:50 pm: Illuminant-adaptive color reproduction for a mobile display, J. M. Kim, K. Park, M. Lee, Y. Cho, Y. Ha, Kyungpook National Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-19] 4:10 pm: Skin color reproduction algorithm for portrait images shown on the mobile display, Y. Kwak, S. Lee, D. Park, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) . . . . . . . . . . . . . [6058-20] 4:30 pm: Estimating displays' color fidelity based on classified image statistics, P. Sun, C. Lee, Shih Hsin Univ. (Taiwan) . . . . . . . . . . . [6058-21] 4:50 pm: End-user display calibration via support vector regression, B. Bastani, Hewlett-Packard Co.; B. V. Funt, W. Xiong, Simon Fraser Univ. (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-22] 5:10 pm: The calibration accuracy of display white point by visual calibrator under various illuminations, T. Sugiyama, Y. Kudo, Dai Nippon Printing Co., Ltd. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-23]
Thursday 19 January SESSION 6 Conv. Ctr. Room A2 . . . . . . . . . . . Thurs. 8:40 to 10:00 am Printing Chair: A. Ufuk Agar, Garanti Technology (Turkey) 8:40 am: Black extraction method using gamut boundary descriptors, M. Cho, B. Kang, H. Choh, SAMSUNG Advanced Institute of Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-24] 9:00 am: Colorimetric characterization based on color correlation in CMYKGO printer, I. Jang, C. Son, T. Park, K. Ko, Y. Ha, Kyungpook National Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-25] 9:20 am: Hardcopy global color correction, Y. Bang, Y. Kim, H. Choh, Samsung Advanced Institute of Technology (South Korea) . . . . . [6058-26] 9:40 am: Efficient document rendering with enhanced run length encoding, G. Feng, Ricoh Innovations, Inc.; C. A. Bouman, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-28] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am SESSION 7 Conv. Ctr. Room A2 . . . . . . . . . . Thurs. 10:30 to 11:50 am Halftoning I Chair: Jan P. Allebach, Purdue Univ. 10:30 am: Model-based clustered-dot screening, S. H. Kim, SAMSUNG Electronics Co., Ltd. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . [6058-29] 10:50 am: AM-FM hybrid color screen design to reduce brightness variation, B. M. Kang, B. T. Ryu, C. Kim, Inha Univ. (South Korea); S. H. Kim, SAMSUNG Electronics Co., Ltd. (South Korea) . . . . . . . . . . [6058-30] 11:10 am: Frequency domain design of cluster dot screens, M. Fischer, D. Shaked, Hewlett-Packard Labs. (Israel) . . . . . . . . . . . . . . . . . . [6058-31] 11:30 am: A spatial domain optimization method to generate plandependent masks, Y. Wu, Hewlett-Packard Co. . . . . . . . . . . . . . [6058-32] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:50 am to 1:30 pm SESSION 8 Conv. Ctr. Room A2 . . . . . . . . . . . . Thurs. 1:30 to 3:30 pm Halftoning II Chair: Reiner Eschbach, Xerox Corp. 1:30 pm: Using errors in halftoning to increase reproduction accuracy, S. Herron, Global Graphics Software Inc. . . . . . . . . . . . . . . . . . . . [6058-33] 1:50 pm: Analysis of misregistration-induced color shifts in the superposition of periodic screens, B. Oztan, G. Sharma, Univ. of Rochester; R. P. Loce, Xerox Corp. . . . . . . . . . . . . . . . . . . . . . . . [6058-35] 2:10 pm: Analysis of color error diffusion with vector error filters, Z. Z. Fan, Xerox Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-36] 2:30 pm: New halftoning method combining the best of masking and error diffusion algorithms, F. Cittadini, Ocй Print Logic Technologies (France) and Univ. Pierre et Marie Curie (France); J. Pervillй, S. Berche, Ocй Print Logic Technologies (France); M. Ben Chouikha, G. Alquiй, Univ. Pierre et Marie Curie (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-37] 2:50 pm: Graph order dither, A. Hausner, Univ. of New Hampshire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-38] 3:10 pm: Optimal halftoning over hexagonal grids, J. Bacca Rodriguez, A. J. Gonzalez Lozano, G. R. Arce, Univ. of Delaware; D. L. Lau, Univ. of Kentucky . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6058-39]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
21
Conference 6059 · Marriott San Carlos Room
Tuesday-Thursday 17-19 January 2006 · Proceedings of SPIE Vol. 6059 Image Quality and System Performance III
Conference Chairs: Luke C. Cui, Lexmark International, Inc.; Yoichi Miyake, Chiba Univ. (Japan) Program Committee: Peter D. Burns, Eastman Kodak Co.; Mark D. Fairchild, Rochester Institute of Technology; Susan Farnand, Eastman Kodak Co.; Frans Gaykema, OCE Technologies BV (Netherlands); Dirk W. Hertel, Polaroid Corp.; Robin B. Jenkin, Cranfield Univ. (United Kingdom); Nathan Moroney, Hewlett-Packard Co.; Rene Rasmussen, Xerox Corp.; Eric K. Zeise, NexPress Solutions, Inc.
Tuesday 17 January
Wednesday 18 January
Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Image quality assessment based on textual structure and noise normalization, C. Zhang, Z. Qiu, Beijing Jiaotong Univ. (China)[6059-36] Quality models for audiovisual streaming, T. C. Thang, Y. S. Kim, C. S. Kim, Y. M. Ro, Information and Communications Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-37] Research on the analysis and measurement of MTF of staring imaging system, D. Lu, Q. Chen, G. Gu, Nanjing Univ. of Science & Technology (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-38] Quantification method of the color break-up phenomena: evaluation of next-generation color wheels for field sequential color displays, J. Thollot, K. Sarayeddine, Thomson R&D France (France); A. Trйmeau, Univ. Jean Monnet Saint-Etienne (France) . . . . . . . . [6059-39] No-reference jerkiness evaluation method for multimedia communications, M. Carli, D. Guida, A. Neri, Univ. degli Studi di Roma Tre (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-40]
Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 1 Marriott San Carlos Room . . . . . . . Wed. 9:30 to 10:10 am Image Quality Understanding Chair: Luke C. Cui, Lexmark International, Inc. 9:30 am: Fundamental questions related to print quality, P. J. Mangin, M. Dubй, Univ. du Quйbec а Trois-Riviиres (Canada) . . . . . . . . . . [6059-01] 9:50 am: What do users really perceive: probing the subjective image quality, G. S. Nyman, J. Radun, T. Leisti, Univ. of Helsinki (Finland); J. Oja, H. J. Ojanen, J. Olives, T. Vuori, Nokia Research Ctr. (Finland); J. P. Hдkkinen, Univ. of Helsinki (Finland) and Nokia Research Ctr. (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-02] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am SESSION 2 Marriott San Carlos Room . . . Wed. 10:40 am to 12:00 pm Perceptual Attributes and Psychometrics Chair: Nathan Moroney, Hewlett-Packard Co. 10:40 am: The effect of image sharpness on quantitative eyemovement data and on image quality evaluation while viewing natural images, T. Vuori, M. Olkkonen, Nokia Corp. (Finland) . . . . . . . . . [6059-03] 11:00 am: Assessing the enhancement of image sharpness, S. Bouzit, Univ. of St. Andrews (United Kingdom); L. W. MacDonald, London College of Communication (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . [6059-04] 11:20 am: Reference-free quality metric using a region-based attention model for JPEG2000 compressed images, R. Barland, A. Saadane, Univ. de Nantes (France) . . . . . . . . . . . . . . . . . . . . . . . [6059-05] 11:40 am: Comparison of various subjective video quality assessment methods, C. Lee, H. Choi, E. Lee, S. Lee, J. Choe, Yonsei Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-06] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:20 pm
22
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6059 · Marriott San Carlos Room
SESSION 3 Marriott San Carlos Room . . . . . . . . Wed. 1:20 to 2:20 pm Perceptual Image Quality Modeling Chair: Susan Farnand, Eastman Kodak Co. 1:20 pm: Selecting significant colors from a complex image for image quality modeling, K. J. Leeming, P. J. Green, London College of Communication (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . [6059-07] 1:40 pm: Comparison of three full-reference color image quality measures, E. Girshtel, V. Slobodyan, J. S. Weissman, A. M. Eskicioglu, The City Univ. of New York . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-08] 2:00 pm: Influence of ambient illumination on adapted and optimal white point, I. Vogels, J. Berentsen, Philips Research Labs. (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-09] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2:20 to 2:30 pm SESSION 4 Marriott San Carlos Room . . . . . . . . Wed. 2:30 to 3:30 pm Objective Attributes Characterization I Chair: Rene S. Rasmussen, Xerox Corp. 2:30 pm: Characterization of digital image noise properties based on RAW data, H. H. Hytti, Tampereen Teknillinen Yliopisto (Finland) [6059-10] 3:00 pm: An evaluation of sharpness in different image displays used for medical imaging, M. Ukishima, T. Nakaguchi, Chiba Univ. (Japan); K. Kato, Canon Inc. (Japan); Y. Fukuchi, Chiba Univ. Hospital (Japan); N. Tsumura, Chiba Univ. (Japan); K. Matsumoto, Canon, Inc.; N. Yanagawa, H. Morita, Chiba Univ. Hospital (Japan); Y. Miyake, Chiba Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-12] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:30 to 4:00 pm
9:40 am: ISO 19751 macro-uniformity, R. S. Rasmussen, Xerox Corp.; K. D. Donohue, Univ. of Kentucky; Y. S. Ng, NexPress Solutions, Inc.; W. C. Kress, Toshiba America DSE; S. Zoltner, Xerox Corp.; F. Gaykema, OCE Technologies BV (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-20] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am SESSION 7 Marriott San Carlos Room . . . . . Thurs. 10:30 to 11:30 am Standardization II: Breakthroughs Chair: Robin B. Jenkin, Cranfield Univ. (United Kingdom) 10:30 am: Edge-raggedness evaluation using slanted-edge analysis, P. D. Burns, Eastman Kodak Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-21] 11:00 am: Statistical interpretation of ISO TC42 dynamic range: risky business, D. R. Williams, P. D. Burns, Eastman Kodak Co. . . . . . [6059-22] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:30 am to 1:10 pm SESSION 8 Marriott San Carlos Room . . . . . . . Thurs. 1:10 to 2:10 pm System Image Quality Characterization and Modeling I Chair: Dirk W. Hertel, Consultant 1:10 pm: The influence of statistical variations on image quality, B. O. Hultgren, D. W. Hertel, Consultant; J. Bullitt, Polaroid Corp. . . . . [6059-25] 1:30 pm: The use of a virtual printer model for the simulation of imaging systems, B. O. Hultgren, Consultant . . . . . . . . . . . . . . . [6059-26] 1:50 pm: Improved pen alignment for bidirectional printing, E. Bernal, J. P. Allebach, Z. Pizlo, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . [6059-27] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2:10 to 2:20 pm
SESSION 5 Marriott San Carlos Room . . . . . . . . Wed. 4:00 to 5:20 pm Objective Attributes Characterization II Chair: Peter D. Burns, Eastman Kodak Co. 4:00 pm: Characterization of printer MTF, W. Jang, Hewlett-Packard Co.; J. P. Allebach, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-13] 4:20 pm: PSF estimation by gradient descent fit to the ESF, E. H. Barney Smith, Boise State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-14] 4:40 pm: Printer banding estimation using the generalized spectrum, N. A. Rawashdeh, I. Shin, K. D. Donohue, Univ. of Kentucky; S. T. Love, Lexmark International, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-15] 5:00 pm: Scanner-based macroscopic color variation estimation, C. Kuo, L. Di, E. K. Zeise, NexPress Solutions, Inc. . . . . . . . . . . . . . [6059-16] Thursday 19 January SESSION 6 Marriott San Carlos Room . . . . . . Thurs. 8:20 to 10:00 am Standardization I: Progess Chair: Eric K. Zeise, NexPress Solutions, Inc. 8:20 am: Viewing conditions, colorimetric measurements, and profile making: the conundrum of standards vs. practical realities, D. Q. McDowell, Standards Consultant . . . . . . . . . . . . . . . . . . . . . . . . . [6059-17] 8:40 am: Progress in digital color workflow understanding in the ICC workflow WG, A. L. McCarthy, Lexmark International, Inc. and International Color Consortium . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-18] 9:00 am: Recent progress in the development of ISO 19751, S. Farnand, Eastman Kodak Co.; E. N. Dalal, Xerox Corp.; Y. S. Ng, NexPress Solutions, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-19]
SESSION 9 Marriott San Carlos Room . . . . . . . Thurs. 2:20 to 3:30 pm System Image Quality Characterization and Modeling II Chair: Frans Gaykema, OCE Technologies BV (Netherlands) 2:20 pm: Further image quality assessment in digital film restoration, M. Chambah, Univ. de Reims Champagne-Ardenne (France); C. Saint Jean, Univ. de La Rochelle (France); F. Helt, Independent Consultant (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-28] 2:50 pm: Development of picture quality monitoring system for IPTV service based on the reduced reference framework, O. Sugimoto, R. Kawada, A. Koike, KDDI R&D Labs. (Japan) . . . . . . . . . . . . . . . . [6059-29] 3:10 pm: An image quality evaluation method for DOE-based camera lenses, S. Lee, Y. Jin, H. Jeong, Samsung Electro-Mechanics Co., Ltd. (South Korea); S. Song, Hanyang Univ. (South Korea) . . . . . . . . . [6059-30] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:30 to 4:00 pm SESSION 10 Marriott San Carlos Room . . . . . . . Thurs. 4:00 to 5:20 pm Image Defects Characterization and Modeling Chair: Yoichi Miyake, Chiba Univ. (Japan) 4:00 pm: Visibility and annoyance of LCD defective subpixels of different colors and surrounds at different positions, H. Ho, J. M. Foley, S. K. Mitra, Univ. of California/Santa Barbara . . . . . . . . . . . . . . . [6059-32] 4:20 pm: Robust detection of defects in imaging arrays, J. Dudas, C. G. Jung, G. H. Chapman, Simon Fraser Univ. (Canada); Z. Koren, I. Koren, Univ. of Massachusetts/Amherst . . . . . . . . . . . . . . . . . . . . . . . . . [6059-33] 4:40 pm: Objective video quality assessment method for freeze distortion based on freeze aggregation, K. Watanabe, J. Okamoto, T. Kurita, NTT Service Integration Labs. (Japan) . . . . . . . . . . . . . . . [6059-34] 5:00 pm: Film grain synthesis and its applications for re-graining, P. Schallauer, R. Mцrzinger, JOANNEUM RESEARCH GmbH (Austria) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6059-35]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
23
Conference 6060 · Conv. Ctr. Room B1
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6060 Visualization and Data Analysis 2006
Conference Chairs: Robert F. Erbacher, Utah State Univ.; Jonathan C. Roberts, Univ. of Kent (United Kingdom); Matti T. Grцhn, CSCScientific Computing Ltd. (Finland); Katy Bцrner, Indiana Univ. Cochairs: Ming C. Hao, Hewlett-Packard Labs.; Pak C. Wong, Pacific Northwest National Lab. Program Committee: Uwe Brinkschulte, Univ. Karlsruhe (Germany); Philip C. Chen, Future, Inc.; L. E. Greenwade, Idaho National Engineering and Environmental Lab.; Hans-Georg Pagendarm, German Aerospace Research Establishment DLR (Germany); Alex T. Pang, Univ. of California/Santa Cruz; Christopher D. Shaw, Georgia Institute of Technology; Kalpathi R. Subramanian, Univ. of North Carolina/ Charlotte; Yinlong Sun, Purdue Univ.; J. E. Swan II, Naval Research Lab.; Craig M. Wittenbrink, NVIDIA; Yingcai Xiao, Univ. of Akron; William J. Yurcik, Univ. of Illinois at Urbana-Champaign
Monday 16 January SESSION 1 Conv. Ctr. Room B1 . . . . . . . . . . . . . Mon. 8:30 to 9:30 am Flow Visualization Chair: Robert F. Erbacher, Utah State Univ. 8:30 am: Multiscale image based flow visualization, A. C. Telea, Technische Univ. Eindhoven (Netherlands); R. Strzodka, Research Ctr. Caesar (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-01] 8:50 am: Visualizing oceanic and atmospheric flows with streamline splatting, Y. Sun, E. Ess, D. Sapirstein, M. Huber, Purdue Univ. . [6060-02] 9:10 am: View-dependent multiresolutional flow texture advection, L. Li, H. Shen, The Ohio State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . [6060-03] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9:30 to 9:40 am SESSION 2 Conv. Ctr. Room B1 . . . . . . . . . . . . Mon. 9:40 to 11:10 am Volume Visualization Chair: Matti T. Grцhn, Ctr. for Scientific Computing (Finland) 9:40 am: Volumetric depth peeling for medical image display, D. M. Borland, J. P. Clarke, J. R. Fielding, R. M. Taylor II, The Univ. of North Carolina at Chapel Hill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-04] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am 10:30 am: Adaptive border sampling for hardware texture-based volume visualization, E. C. LaMar, Institute for Scientific Research, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-05] 10:50 am: Ray-casting time-varying volume data sets with frame-toframe coherence, D. Tost, S. Grau, Univ. Politиcnica de Catalunya (Spain); M. Ferre, Univ. Rovira i Virgili (Spain); A. Puig, Univ. de Barcelona (Spain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-06] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:10 to 11:20 am SESSION 3 Conv. Ctr. Room B1 . . . . . . . . Mon. 11:20 am to 12:00 pm Visualization Theory Chair: Pak C. Wong, Pacific Northwest National Lab. 11:20 am: Theoretical analysis of uncertainty visualizations, T. D. Zuk, M. S. T. Carpendale, Univ. of Calgary (Canada) . . . . . . . . . . . . . . [6060-07] 11:40 am: A visualization framework for design and evaluation, B. J. Blundell, S. Pettifer, The Univ. of Manchester (United Kingdom); G. Ng, Cerebra, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-08] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm
SESSION 4 Conv. Ctr. Room B1 . . . . . . . . . . . . . Mon. 1:30 to 2:00 pm Invited Paper I Chair: Robert F. Erbacher, Utah State Univ. 1:30 pm: Visual analytics and the NVAC (Invited Paper, Presentation Only), P. C. Wong, Pacific Northwest National Lab. . . . . . . . . . . . [6060-09] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2:00 to 2:10 pm SESSION 5 Conv. Ctr. Room B1 . . . . . . . . . . . . . Mon. 2:10 to 3:10 pm Lighting Chair: Matti T. Grцhn, Ctr. for Scientific Computing (Finland) 2:10 pm: Maximum entropy lighting for physical objects, T. Malzbender, E. Ordentlich, Hewlett-Packard Labs. . . . . . . . . . . . . . . . . . . . . . [6060-10] 2:30 pm: Pre-computed illumination for isosurfaces, K. M. Beason, Florida State Univ.; J. Grant, Pixar Animation Studios; D. C. Banks, B. Futch, M. Y. Hussaini, Florida State Univ. . . . . . . . . . . . . . . . . . . . [6060-11] 2:50 pm: Retro-rendering with vector-valued light: producing local illumination from the transport equation, D. C. Banks, K. M. Beason, Florida State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-12] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm SESSION 6 Conv. Ctr. Room B1 . . . . . . . . . . . . . Mon. 3:40 to 4:20 pm Image Processing Chair: Ming C. Hao, Hewlett-Packard Labs. 3:40 pm: Bit-plane based analysis of integer wavelet coefficients for image compression, A. F. Abu-Hajar, Digitavid, Inc. . . . . . . . . . . [6060-13] 4:00 pm: Two-dimensional reduction PCA: a novel approach for feature extraction, representation, and recognition, R. M. Mutelo, W. L. Woo, S. S. Dlay, Univ. of Newcastle Upon Tyne (United Kingdom) . . . . . . [6060-14] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4:20 to 4:30 pm SESSION 7 Conv. Ctr. Room B1 . . . . . . . . . . . . . Mon. 4:30 to 5:10 pm Terrain/GIS Visualization Chair: Edward Suvanaphen, Univ. of Kent (United Kingdom) 4:30 pm: Energetically optimal travel across terrain: visualizations and a new metric of geographic distance with archaeological applications, B. Wood, Harvard Univ.; Z. Wood, California Polytechnic State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-15] 4:50 pm: Real-time 3D visualization of DEM combined with a robust DCT-based data-hiding method, A. Martin, Univ. Montpellier II (France); G. Gesquiere, Univ. de Provence (France); W. Puech, Univ. Montpellier II (France); S. Thon, Univ. de Provence (France) . . . . . . . . . . . . . . . [6060-16] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:10 to 5:20 pm
24
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6060 · Conv. Ctr. Room B1
SESSION 8 Conv. Ctr. Room B1 . . . . . . . . . . . . . Mon. 5:20 to 6:00 pm Applications Chair: Katy Bцrner, Indiana Univ. 5:20 pm: Hierarchical causality explorer: making complemental use of 3D/2D visualizations, S. Azuma, Ochanomizu Univ. (Japan); I. Fujishiro, Tohoku Univ. (Japan); H. Horii, The Univ. of Tokyo (Japan) . . . . . [6060-17] 5:40 pm: InvIncrements: incremental software to support visual simulation, D. C. Banks, W. Blanco, Florida State Univ. . . . . . . . [6060-18] Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 9 Conv. Ctr. Room B1 . . . . . . . . . . . . Tues. 9:40 to 10:20 am Interaction Techniques Chair: Ming C. Hao, Hewlett-Packard Labs. 9:40 am: Plot of plots and selection glass, H. Chen, SAS Institute Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-19] 10:00 am: Navigation techniques for large-scale astronomical exploration, C. Fu, The Hong Kong Univ. of Science and Technology (Hong Kong China); A. J. Hanson, E. A. Wernert, Indiana Univ. . . [6060-20] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am SESSION 10 Conv. Ctr. Room B1 . . . . . . . . . . . Tues. 10:50 to 11:50 am InfoVis Chair: Ketan K. Mane, Indiana Univ. 10:50 am: Reducing InfoVis cluttering through non uniform sampling, displacement, and user perception, E. Bertini, G. Santucci, L. Dell'Aquila, Univ. degli Studi di Roma/La Sapienza (Italy) . . . . . . [6060-21] 11:10 am: Diverse information integration and visualization, S. L. Havre, A. R. Shah, C. Posse, B. M. Webb-Robertson, Pacific Northwest National Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-22] 11:30 am: WordSpace: visual summary of text corpora, U. Brandes, M. Hoefer, J. Lerner, Univ. of Konstanz (Germany) . . . . . . . . . . . . . . [6060-23] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:50 am to 1:40 pm SESSION 11 Conv. Ctr. Room B1 . . . . . . . . . . . . . Tues. 1:40 to 2:10 pm Invited Paper II Chair: Ketan K. Mane, Indiana Univ. 1:40 pm: Information architecture: why design matters (Invited Paper, Presentation Only), J. Agutter, Univ. of Utah . . . . . . . . . . . . . . . [6060-24] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2:10 to 2:20 pm SESSION 12 Conv. Ctr. Room B1 . . . . . . . . . . . . . Tues. 2:20 to 3:20 pm Visualization Techniques I Chair: Matti T. Grцhn, Ctr. for Scientific Computing (Finland) 2:20 pm: Trees in a treemap: visualizing multiple hierarchies, M. Burch, S. Diehl, Katholische Univ. Eichstдtt (Germany) . . . . . . . . . . . . . . [6060-25]
2:40 pm: Focus-based filtering + clustering technique for power-law networks with small world phenomenon, F. Boutin, J. Thiиvre, M. Hascoлt, Univ. Montpellier II (France) and CNRS (France) . . . . . . [6060-26] 3:00 pm: Enhancing scatterplot matrices for data with ordering or spatial attributes, Q. Cui, M. O. Ward, E. A. Rundensteiner, Worcester Polytechnic Institute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-27] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:20 to 3:50 pm SESSION 13 Conv. Ctr. Room B1 . . . . . . . . . . . . . Tues. 3:50 to 4:30 pm Visualization Techniques II Chair: Edward Suvanaphen, Univ. of Kent (United Kingdom) 3:50 pm: Content-based text mapping using multidimensional projections for exploration of document collections, R. Minghim, F. V. Paulovich, A. de Andrade Lopes, Univ. de Sгo Paulo (Brazil) . . . [6060-28] 4:10 pm: Mapping texts through dimensionality reduction and visualization techniques for interactive exploration of document collections, A. d. A. de Andrade Lopes, R. Minghim, V. Melo, F. V. Paulovich, Univ. de Sгo Paulo (Brazil) . . . . . . . . . . . . . . . . . . . . . [6060-29] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4:30 to 4:40 pm SESSION 14 Conv. Ctr. Room B1 . . . . . . . . . . . . . Tues. 4:40 to 5:20 pm Bioinformatics Chair: Robert F. Erbacher, Utah State Univ. 4:40 pm: Visualizing brain rhythms and synchrony, K. A. Robbins, D. Veljkovic, E. Pilipaviciute, The Univ. of Texas at San Antonio . . . . [6060-30] 5:00 pm: Automatic feature-based surface mapping for brain cortices, L. Linsen, Ernst Moritz Arndt Univ. Greifswald (Germany) . . . . . . [6060-31] Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Blogviz: mapping the dynamics of information diffusion in blogspace, M. S. Lima, Parsons School of Design . . . . . . . . . [6060-32] Organizing and visualizing database data using parallel coordinates, C. G. Presser, Gettysburg College . . . . . . . . . . . [6060-33] Visualizing 3D vector fields with splatted streamlines, E. Ess, Y. Sun, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-35] SRS browser: a visual interface to the sequence retrieval system, K. K. Mane, K. Borner, Indiana Univ. . . . . . . . . . . . . . . . . . . . . . . . [6060-36] Tracing parallel vectors, J. Sukharev, Univ. of California/Santa Cruz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-39] Output-sensitive volume tracking, L. Jiang, Rutgers Univ. . . [6060-40] Visualization of force fields in protein structure prediction, S. N. Crivelli, Lawrence Berkeley National Lab. and California Institute for Quantitative Biomedical Research; C. Crawford, O. Kreylos, B. Hamann, Univ. of California/Davis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6060-41] Correspondence-based visualization techniques, M. J. GeraldYamasaki, NASA Ames Research Ctr. . . . . . . . . . . . . . . . . . . . [6060-42]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
25
Conference 6061 · Conv. Ctr. Room B4
Wednesday-Thursday 18-19 January 2006 · Proceedings of SPIE Vol. 6061 Internet Imaging VII Conference Chairs: Simone Santini, Univ. of California/San Diego; Raimondo Schettini, DISCo/Univ. degli Studi di Milano-Bicocca (Italy); Theo Gevers, Univ. van Amsterdam (Netherlands) Program Committee: Jeffrey E. Boyd, Univ. of Calgary (Canada); Alberto Del Bimbo, Univ. degli Studi di Firenze (Italy); Jennifer Gille, Raytheon Co.; Hagit Z. Hel-Or, Univ. of Haifa (Israel); Roger David Hersch, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland); Yasuyo G. Ichihara, Hosen-Gakuen College (Japan); Reiner Lenz, Linkцpings Univ. (Sweden); Clement H. C. Leung, Victoria Univ. of Technology (Australia); Yong Rui, Microsoft Research; Simon Shim, San Josй State Univ.; Alain Trйmeau, Univ. Jean Monnet (France); Luc J. Van Gool, Katholieke Univ. Leuven (Belgium)
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Subjective trajectory characterization: acquisition, matching, and retrieval, M. Y. Zhang, L. Olsen, J. E. Boyd, Univ. of Calgary . [6061-30] Archiving of meaningful scenes for personal TV terminals, S. H. Jin, J. H. Cho, Y. M. Ro, Information and Communications Univ. (South Korea); J. Kim, Electronics and Telecommunications Research Institute (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-32] AVIR: a spoken document retrieval system in e-learning environment, I. Gagliardi, M. Padula, P. Pagliarulo, Consiglio Nazionale delle Ricerche (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-33] Internet-based remote counseling to support stress management: preventing interruptions to regular exercise in elderly people, S. Hashimoto, Univ. of Tsukuba (Japan) and National Institute of Information and Communications Technology (Japan); T. Munakata, Univ. of Tsukuba (Japan); N. Hashimoto, Citizen Watch Co., Ltd. (Japan); J. Okunaka, T. Koga, National Institute of Information and Communications Technology (Japan) . . . . . . . . . . . . . . . . . . . [6061-34] Vertex and face permutation order compression for efficient animation support, E. Chang, Electronics and Telecommunications Research Institute (South Korea); D. Kim, B. Min, S. Lee, Hanyang Univ. (South Korea); N. Hur, S. Lee, Electronics and Telecommunications Research Institute (South Korea); E. S. Jang, Hanyang Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-35]
Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 1 Conv. Ctr. Room B4 . . . . . . . . . . . . Wed. 9:30 to 10:20 am Special Session: Benchmarking I 9:30 am: Requirements for benchmarking personal image retrieval systems (Invited Paper), J. Bouguet, C. Dulong, I. V. Kozintsev, Intel Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-01] 10:00 am: On usage models of content-based image search, filtering, and annotation, D. Telleen-Lawton, C. B. Chang, VIMA Technologies, Inc.; E. Y. Chang, Univ. of California/Santa Barbara . . . . . . . . . . . . . . . [6061-02] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am SESSION 2 Conv. Ctr. Room B4 . . . . . . . . . . . Wed. 10:50 to 11:50 am Special Session: Benchmarking II 10:50 am: Human factors in automatic image retrieval system design and evaluation, A. Jaimes, Fuji Xerox Co., Ltd. (Japan) . . . . . . . [6061-03] 11:10 am: Lessons from TRECVID: lexicon design for semantic indexing in media databases, M. R. Naphade, IBM Thomas J. Watson Research Ctr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-04] 11:30 am: Benchmarking without ground truth, S. Santini, Univ. of California/San Diego . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-21] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:30 am to 1:00 pm SESSION 3 Conv. Ctr. Room B4 . . . . . . . . . . . . . Wed. 1:00 to 2:00 pm Special Session: Benchmarking III 1:00 pm: Using heterogeneous annotation and visual information for the benchmarking of image retrieval system, H. Mьller, Univ. Hospital of Geneva (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-06] 1:20 pm: On benchmarking content-based image retrieval applications, B. Zhang, Y. Zuo, Tsinghua Univ. (China) . . . . . . . . [6061-07] 1:40 pm: TRECVID: the utility of a content-based video retrieval evaluation, A. G. Hauptmann, Carnegie Mellon Univ. . . . . . . . . . [6061-08]
26
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6061 · Conv. Ctr. Room B4
SESSION 4 Conv. Ctr. Room B4 . . . . . . . . . . . . . Wed. 2:00 to 3:10 pm Interfaces and Visualization 2:00 pm: A color selection tool ensuring legibility of textual information on web pages (Invited Paper), S. Zuffi, Consiglio Nazionale delle Ricerche (Italy); G. B. Beretta, Hewlett-Packard Co.; C. Brambilla, Consultant (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-10] 2:30 pm: A color interface for audio clustering visualization, S. Zuffi, I. Gagliardi, Consiglio Nazionale delle Ricerche (Italy) . . . . . . . . . . . [6061-11] 2:50 pm: Interactive internet delivery of scientific visualization vis structured prerendered imagery, J. Chen, San Francisco State Univ. and Lawrence Berkeley National Lab.; E. W. Bethel, Lawrence Berkeley National Lab.; I. Yoon, San Francisco State Univ. . . . . . . . . . . . . [6061-12] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm SESSION 5 Conv. Ctr. Room B4 . . . . . . . . . . . . . Wed. 3:40 to 4:20 pm Ontology and Annotation 3:40 pm: Clustering and semantically filtering web images to create a large-scale image ontology, S. Zinger, C. Millet, M. Benoit, G. Grefenstette, P. Hиde, P. Moлllic, Commissariat a l'Energie Atomique (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-13] 4:00 pm: Ontology and image semantics in multimodal imaging: submission and retrieval, Y. Bei, M. Belmamoune, F. J. Verbeek, Leiden Univ. (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-14] Thursday 19 January
SESSION 8 Conv. Ctr. Room B4 . . . . . . . . . . . . Thurs. 1:30 to 2:10 pm Video 1:30 pm: Enhanced video display and navigation for networked streaming video and networked video playlists, S. G. Deshpande, Sharp Labs. of America, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-24] 1:50 pm: 3D display technique for moving pictures from web cameras using screen pixel accessing, T. Hasegawa, T. Namiki, H. Unno, K. Uehira, H. Kasuga, K. Yanaka, Kanagawa Institute of Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-25] SESSION 9 Conv. Ctr. Room B4 . . . . . . . . . . . . Thurs. 2:10 to 2:50 pm Vector Displays 2:10 pm: Dynamic conversion between XML-based languages for vector graphics, A. Di Iorio, F. Vitali, G. Zonta, Univ. degli Studi di Bologna (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-28] 2:30 pm: Bezier curves approximation of triangularized surfaces using SVG, G. Messina, STMicroelectronics (Italy); E. Ingra, S. Battiato, G. Di Blasi, Univ. di Catania (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-29]
SESSION 6 Conv. Ctr. Room B4 . . . . . . . . . . . Thurs. 9:00 to 10:00 am Anthropometrics 9:00 am: Combining color models for skin detection, F. Aldershoff, T. Gevers, H. M. Stokman, Univ. van Amsterdam (Netherlands) . . . [6061-16] 9:20 am: Using context and similarity for face and location identification, M. Davis, Univ. of California/Berkeley; M. A. Smith, France Telecom R&D (France); J. F. Canny, Univ. of California/Berkeley; F. W. M. Stentiford, Univ. College London (United Kingdom) . . . . . . . . . . . [6061-17] 9:40 am: Skin segmentation using multiple thresholding, F. Gasparini, R. Schettini, Univ. degli Studi di Milano-Bicocca (Italy) . . . . . . . . [6061-18] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am
SESSION 7
Conv. Ctr. Room B4 . . . . . . . Thurs. 10:30 am to 12:10 pm content management and Retrieval 10:30 am: Integration of multimedia contents and e-learning resources in a digital library, M. Pascual, N. Ferran, J. Minguillуn Alfonso, Univ. Oberta de Catalunya (Spain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-19] 10:50 am: Selecting the kernel type for a web-based adaptive image retrieval systems (AIRS), A. Doloc-Mihu, V. V. Raghavan, Univ. of Louisiana at Lafayette . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-20] 11:10 am: FaceLab: a tool for performance evaluation of face recognition strategies, L. Caflisch, Comerson s.r.l. (Italy); A. Colombo, C. Cusano, R. Schettini, F. Tisato, Univ. degli Studi di Milano-Bicocca (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-36] 11:30 am: Medical validation and CBIR of spine x-ray images over the Internet, S. K. Antani, J. Cheng, J. L. Long, L. R. Long, G. R. Thoma, National Library of Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-22] 11:50 am: The integration of cartographic information into a content management system, M. M. Furnari, C. I. D. Noviello, Consiglio Nazionale delle Ricerche (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6061-23] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:30 pm
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
27
Conference 6062 · Conv. Ctr. Room A2
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6062 Spectral Imaging: Eighth International Symposium on Multispectral Color Science Conference Chairs: Mitchell R. Rosen, Rochester Institute of Technology; Francisco H. Imai, Pixim, Inc.; Shoji Tominaga, Osaka ElectroCommunication Univ. (Japan) Program Committee: Roy S. Berns, Rochester Institute of Technology; Jeffrey M. DiCarlo, Hewlett-Packard Labs.; Jon Y. Hardeberg, Gjшvik Univ. College (Norway); Markku Hauta-Kasari, Univ. of Joensuu (Finland); Bernhard Hill, Univ. Aachen (Germany); Reiner Lenz, Linkцping Univ. (Sweden); Yoshitsugu Manabe, Nara Institute of Science and Technology (Japan); Yoichi Miyake, Chiba Univ. (Japan); Javier Romero, Univ. de Granada (Spain); Norimichi Tsumura, Chiba Univ. (Japan); Stephen Westland, Univ. of Derby (United Kingdom)
Monday 16 January SESSION 1 Conv. Ctr. Room A2 . . . . . . . . . . . . Mon. 9:00 to 10:20 am Spectral Analysis for Scene Content Identification I Chair: Francisco H. Imai, Rochester Institute of Technology 9:00 am: Hyperspectral imaging of sulfate evaporite deposits in Western Australia and on Mars (Invited Paper), A. J. Brown, Macquarie Univ. (Australia) and Australian Ctr. for Astrobiology (Australia); T. J. Cudahy, Commonwealth Scientific & Industrial Research Organisation (Australia) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-01] 9:40 am: Multispectral imaging determination of pigment concentration profiles in meat, C. Sбenz, B. Hernбndez, C. Alberdi, S. Alfonso, M. Berrogui, J. M. Diсeiro, Univ. Publica de Navarra (Spain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-02] 10:00 am: Visualization of the human face skin moisturizing-ability by spectroscopic imaging using two near-infrared bands, H. Iwasaki, K. Miyazawa, S. Nakauchi, Toyohashi Univ. of Technology (Japan) . [6062-03] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am SESSION 2 Conv. Ctr. Room A2 . . . . . . . . Mon. 10:50 am to 12:20 pm Spectral Analysis for Scene Content Identification II Chair: Javier Hernandez-Andres, Univ. de Granada (Spain) 10:50 am: Spectral estimation of made-up skin color under various conditions (Invited Paper), M. Doi, R. Ohtsuki, S. Tominaga, Osaka Electro-Communication Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . [6062-05] 11:20 am: Image processing techniques for detection of buried objects in infrared images, A. Ceron-Correa, Univ. Militar Nueva Granada (Colombia); O. L. Lopera, Royal Military Academy (Belgium) and Univ. de Los Andes (Colombia) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-04] 11:40 am: MODIS versus ASTER water classification, C. Alecu, S. Oancea, National Meteorological Administration (Romania); E. Bryant, Dartmouth College . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-06] 12:00 pm: Improving multivariate curve resolution analysis performance when applied to fluorescence hyperspectral biological imaging (Presentation Only), H. D. T. Jones, E. Thomas, D. M. Haaland, J. A. Timlin, M. B. Sinclair, Sandia National Labs. . . . . . . . . . . . . [6062-07] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:20 to 1:50 pm
SESSION 3 Conv. Ctr. Room A2 . . . . . . . . . . . . . Mon. 1:50 to 2:30 pm Spectral Estimation Methods Chair: Masahiro Yamaguchi, Tokyo Institute of Technology (Japan) 1:50 pm: Estimating reflectance parameters from saturated spectral images, S. Li, Y. Manabe, K. Chihara, Nara Institute of Science and Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-08] 2:10 pm: Influence of the recovery method in the optimum sensors for spectral imaging of skylight, M. A. Lopez-Alvarez, J. Hernandez-Andres, J. L. Nieves, J. Romero, Univ. de Granada (Spain) . . . . . . . . . . . . [6062-09] SESSION 4 Conv. Ctr. Room A2 . . . . . . . . . . . . . Mon. 2:30 to 4:50 pm Spectral Acquisition Systems Chairs: Javier Romero, Univ. de Granada (Spain); Markku Hauta- Kasari, Joensuu Yliopisto (Finland) 2:30 pm: Demosaicking methods for multispectral cameras using mosaic focal plane array technology (Invited Paper), G. A. Baone, H. Qi, The Univ. of Tennessee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-10] 3:00 pm: Estimation of noise variance of a multispectral image acquisition system, N. Shimano, Kinki Univ. (Japan) . . . . . . . . . [6062-11] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:20 to 3:50 pm 3:50 pm: Multispectral stand-off imaging with mid-infrared semiconductor lasers, Y. Wang, Y. Wang, H. Q. Le, Univ. of Houston . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-12] 4:10 pm: Designing flat-bed scanning system for spectral and glossiness recording, T. Takiguchi, S. Abe, T. Makino, N. Tsumura, T. Nakaguchi, Chiba Univ. (Japan); F. Nakaya, H. Ichikawa, Y. Minato, Fuji Xerox Co., Ltd. (Japan); K. Miyata, National Museum of Japanese History (Japan); Y. Miyake, Chiba Univ. (Japan) . . . . . . . . . . . . . . . . . . . . [6062-13] 4:30 pm: Color measurements with colorimetric and multispectral imaging systems, M. de Lasarte, M. Vilaseca, J. Pujol, M. Arjona, Univ. Politиcnica de Catalunya (Spain) . . . . . . . . . . . . . . . . . . . . . . . . . [6062-15]
28
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6062 · Conv. Ctr. Room A2
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 5 Conv. Ctr. Room A2 . . . . . . . . . Tues. 9:30 am to 12:00 pm Spectral Reproduction Chair: Shoji Tominaga, Osaka Electro-Communication Univ. (Japan) 9:30 am: High-fidelity video and still-image communication based on spectral information: natural vision system and its applications (Invited Paper), M. Yamaguchi, Tokyo Institute of Technology (Japan); H. Haneishi, Chiba Univ. (Japan); H. Fukuda, J. Kishimoto, H. Kanazawa, M. Tsuchida, R. Iwama, National Institute of Information and Communications Technology (Japan); N. Ohyama, Tokyo Institute of Technology ( Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-16] 10:00 am: Encoding of spectra for multiple observers and multiple illuminants, T. Boosmann, RWTH Aachen (Germany) . . . . . . . . . [6062-17] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am 10:50 am: Spectral-based color reproduction for print illuminated by image projector (Invited Paper), K. Ueda, S. Yamamoto, N. Tsumura, T. Nakaguchi, Y. Miyake, Chiba Univ. (Japan) . . . . . . . . . . . . . . . . . [6062-18] 11:20 am: Spectral-based optimization of screen images for industrial product presentation, L. Hдrkцnen, J. B. Martinkauppi, H. T. Laamanen, M. Hauta-Kasari, Joensuu Yliopisto (Finland); P. Huhtelin, P. Horttanainen, Tulikivi Oyj (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-19] 11:40 am: Spectral gamuts and spectral gamut mapping, M. R. Rosen, Rochester Institute of Technology; M. W. Derhak, Onyx Graphics [6062-20] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm SESSION 6 Conv. Ctr. Room A2 . . . . . . . . . . . . . Tues. 1:30 to 2:20 pm Art Spectral Imaging Chair: Norimichi Tsumura, Chiba Univ. (Japan) 1:30 pm: A technique for detecting metameric color areas for investigation of historical materials (Invited Paper), K. Miyata, National Museum of Japanese History (Japan); H. T. Laamanen, T. Jaaskelainen, M. Hauta-Kasari, J. P. Parkkinen, Joensuu Yliopisto (Finland) . . . . . [6062-21] 2:00 pm: A scanning device for multispectral imaging of paintings, C. Bonifazzi, Univ. degli Studi di Ferrara (Italy); P. Carcagnм, A. D. Patria, Istituto Nazionale di Ottica Applicata (Italy); S. Ferriani, ENEA (Italy); R. Fontana, M. Greco, M. G. Mastroianni, M. Materazzi, E. M. Pampaloni, A. Romano, Istituto Nazionale di Ottica Applicata (Italy) . . . . . . . . . [6062-22]
SESSION 7 Conv. Ctr. Room A2 . . . . . . . . . . . . . Tues. 2:20 to 3:00 pm Spectral Video Systems Chair: Mitchell R. Rosen, Rochester Institute of Technology 2:20 pm: Spectral video intraframe compression and database, J. P. Purmonen, M. Hauta-Kasari, J. Tuomela, Joensuu Yliopisto (Finland); M. Yamaguchi, M. Mitsui, Tokyo Institute of Technology (Japan); H. Fukuda, National Institute of Information and Communications Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-23] 2:40 pm: Real-rime, multispectral, color acquisition and display using commodity hardware components, D. L. Lau, A. M. Tan, Univ. of Kentucky . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-24] Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Construction of multichannel camera gamuts, S. Helling, RWTH Aachen (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-25] Importance of the texture features in a query from spectral image databases, O. Kohonen, M. Hauta-Kasari, Univ. of Joensuu (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6062-27]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
29
Conference 6063 · Conv. Ctr. Room B3
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6063 Real-Time Image Processing III
Conference Chairs: Nasser Kehtarnavaz, The Univ. of Texas at Dallas; Phillip A. Laplante, The Pennsylvania State Univ. Program Committee: Mohamed Akil, Йcole Supйrieure d'Ingйnieurs en Йlectronique et Йlectrotechnique (France); Matthias F. Carlsohn, Computer Vision and Image Communication (Germany); Carlos R. Castro-Pareja, Univ. of Maryland/Baltimore; Luciano F. da Fontoura Costa, Univ. de Sгo Paulo (Brazil); Philip P. Dang, STMicroelectronics; Xavier Desurmont, Multitel (Belgium); Edward R. Dougherty, Texas A&M Univ.; Sang-Yong Lee, Texas Instruments Inc.; Chang-Joon Park, Electronics and Telecommunications Research Institute (South Korea); Gregory Pisanich, NASA Ames Research Ctr.; Volodymyr I. Ponomaryov, Instituto Politйcnico Nacional (Mexico); Fatih M. Porikli, Mitsubishi Electric Research Labs.; Raghvinder S. Sangwan, The Pennsylvania State Univ.; Feng Xiao, Agilent Technologies
Monday 16 January SESSION 1 Conv. Ctr. Room B3 . . . . . . . . . . . . Mon. 8:50 to 10:00 am Medical Applications Chair: Nasser Kehtarnavaz, The Univ. of Texas at Dallas 8:30 am: Fast computation of free-form deformations in 3D images using FPGAs (Invited Paper), C. R. Castro-Pareja, Intel Corp.; R. Shekhar, Univ. of Maryland/Baltimore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-01] 9:00 am: Toward real-time stereoscopic depth reconstruction in laparoscopic surgery, B. J. McCullagh, F. P. Shevlin, The Univ. of Dublin, Trinity College (Ireland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-02] 9:20 am: Real-time wavelet denoising with edge enhancement for medical x-ray imaging, G. Luo, D. Osypiw, Buckinghamshire Chilterns Univ. College (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-03] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am SESSION 2 Conv. Ctr. Room B3 . . . . . . . . Mon. 10:30 am to 12:00 pm Video Processing Chair: Carlos R. Castro-Pareja, Intel Corp. 10:30 am: Real-time high-level video understanding using data warehouse (Invited Paper), B. Lienard, X. Desurmont, B. Barrie, J. Delaigle, Multitel A.S.B.L. (Belgium) . . . . . . . . . . . . . . . . . . . . . . . [6063-05] 11:00 am: Video surveillance using distance maps, T. E. Schouten, H. C. Kuppens, Radboud Univ. Nijmegen (Netherlands); E. L. van den Broek, Vrije Univ. Amsterdam (Netherlands) and Katholieke Univ./Nijmegen (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-06] 11:20 am: Vehicle traffic video data real-time processing, M. Andreae, W. K. Cheng, Massachusetts Institute of Technology . . . . . . . . . [6063-07] 11:40 am: Vehicle counting system using real-time video processing, P. Crisostomo-Romero, Pontificia Univ. Catolica del Peru (Peru) . . . [6063-08] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm
SESSION 3 Conv. Ctr. Room B3 . . . . . . . . . . . . . Mon. 1:30 to 3:00 pm Algorithms Chair: Xavier Desurmont, Multitel A.S.B.L. (Belgium) 1:30 pm: Real-time auto white balancing using DWT-based multiscale clustering (Invited Paper), N. Kehtarnavaz, N. Kim, M. N. Gamadia, The Univ. of Texas at Dallas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-09] 2:00 pm: Real-time antialiasing using adaptive directional filtering, P. Rokita, Politechnika Warszawska (Poland) . . . . . . . . . . . . . . . . . . [6063-10] 2:20 pm: A fast eye detector using corners, color, and edges, L. Chen, C. Grecos, Loughborough Univ. (United Kingdom) . . . . . . . . . . . [6063-11] 2:40 pm: Real-time construction of covariance matrices for arbitrary size image windows, F. M. Porikli, O. Tuzel, Mitsubishi Electric Research Labs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-12] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm SESSION 4 Conv. Ctr. Room B3 . . . . . . . . . . . . . Mon. 3:30 to 5:40 pm Hardware Chair: Mohamed Akil, Йcole Supйrieure d'Ingйnieurs en Йlectronique et Йlectrotechnique (France) 3:30 pm: High-performance VLSI architecture for adaptive scaling (Invited Paper), P. P. Dang, STMicroelectronics . . . . . . . . . . . . . [6063-13] 4:00 pm: Architecture for hardware driven image inspection based on FPGAs, J. Fuertler, J. Brodersen, Austrian Research Ctrs. GmbH Seibersdorf (Austria); P. Roessler, Technische Univ. Wien (Austria); K. J. Mayer, Austrian Research Ctrs. GmbH - Seibersdorf (Austria); G. Cadek, C. Eckel, H. Nachtnebel, Technische Univ. Wien (Austria) . . . . . . . . [6063-14] 4:20 pm: Using a field programmable object array (FPOA) to accelerate image processing, S. Riley, MathStar, Inc. . . . . . . . [6063-15] 4:40 pm: Novel windowing technique realized in FPGA for radar system, E. Escamilla-Hernбndez, Instituto Politйcnico Nacional (Mexico); V. F. Kravchenko, Instituto Politйcnico Nacional (Mexico) and Institute of Radio Engineering and Electronics (Russia); V. I. Ponomaryov, Instituto Politйcnico Nacional (Mexico) . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-16] 5:00 pm: Real-time hardware for a new 3D display, M. Akil, B. Kaufmann, Йcole Supйrieure d'Ingйnieurs en Йlectronique et Йlectrotechnique (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-17] 5:20 pm: A rapid prototyping methodology to implement and optimize image processing algorithms for FPGAs, M. Akil, P. Niang, T. Grandpierre, Йcole Supйrieure d'Ingйnieurs en Йlectronique et Йlectrotechnique (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-18]
30
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6063 · Conv. Ctr. Room B3 Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. An efficient illuminance-reflectance nonlinear video stream enhancement model, L. Tao, V. K. Asari, Old Dominion Univ. [6063-19] A novel two-pass hexagonal search algorithm for motion estimation, Y. Wu, The Univ. of Reading (United Kingdom) . . [6063-21] Real-time image processing based on robust linear combinations of order statistics, F. J. Gallegos-Funes, J. L. Varela-Benitez, V. I. Ponomaryov, Instituto Politйcnico Nacional (Mexico) . . . . . . . [6063-22] A new concept of real-time security camera monitoring with privacy protection by masking moving objects, K. Yabuta, H. Kitazawa, T. Tanaka, Tokyo Univ. of Agriculture and Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-23] Online monitoring for wood pieces on a moving conveyor belt, W. Wang, Chongqing Univ. of Posts and Telecommunications (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-24] A hardware-accelerated approach to computing multiple image similarity measures from joint histogram, C. R. Castro-Pareja, Intel Corp.; R. Shekhar, Univ. of Maryland/Baltimore . . . . . . . . . . . . [6063-25] Real-time human detection by shape and motion, H. Ran, Wuhan Univ. of Technology (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-26] Uranus: an environment for rapid prototyping of real-time video processing based on FPGA, M. A. Nuсo-Maganda, V. H. RosalesHernбndez, L. N. Castillo-Jimenez, G. Sosa-Ramнrez, M. O. AriasEstrada, Instituto Nacional de Astrofнsica, Уptica y Electrуnica (Mexico) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-31] Determination of traffic intensity from camera images using image processing and pattern recognition techniques, M. Mehrubeoglu, Texas A&M Univ./Corpus Christi; L. McLauchlan, Texas A&M Univ./ Kingsville . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6063-32]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
31
Conference 6064A · Conv. Ctr. Room C2
Monday-Tuesday 16-17 January 2006 · Part of Proceedings of SPIE Vol. 6064 Image Processing: Algorithms and Systems, Neural Networks and Machine Learning Image Processing: Algorithms and Systems V
Conference Chairs: Edward R. Dougherty, Texas A&M Univ.; Jaakko T. Astola, Tampere Univ. of Technology (Finland); Karen O. Egiazarian, Tampere Univ. of Technology (Finland) Program Committee: Til Aach, RWTH Aachen (Germany); Sos S. Agaian, The Univ. of Texas at San Antonio; Junior Barrera, Univ. de Sгo Paulo (Brazil); Reiner Creutzburg, Fachhochschule Brandenburg (Germany); Paul D. Gader, Univ. of Florida; Atanas P. Gotchev, Tampere Univ. of Technology (Finland); John C. Handley, Xerox Corp.; Vladimir V. Lukin, National Aerospace Univ. (Ukraine); Stephen Marshall, Univ. of Strathclyde (United Kingdom); Franзoise J. Prкteux, Institut National des Tйlйcommunications (France); Giovanni Ramponi, Univ. Degli Studi di Trieste (Italy); Jagath K. Samarabandu, The Univ. of Western Ontario (Canada); Akira Taguchi, Musashi Institute of Technology (Japan)
Monday 16 January SESSION 1 Conv. Ctr. Room C2 . . . . . . . . . Mon. 9:00 am to 12:10 pm Image Processing Algorithms 9:00 am: Affine invariant surface evolutions for 3D image segmentation, Y. Rathi, Georgia Institute of Technology; P. Olver, G. Sapiro, Univ. of Minnesota; A. R. Tannenbaum, Georgia Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-01] 9:20 am: Iterative Markovian estimation of mass functions in Dempster Shafer evidence theory: application to multisensor image segmentation, L. Bentabet, M. Jiang, Bishop's Univ. (Canada) [6064A-02] 9:40 am: Progressive halftoning by Perona-Malik error diffusion and stochastic flipping, J. J. Shen, Univ. of Minnesota . . . . . . . . . [6064A-03] 10:00 am: Edge-based stochastic active contours for medical imaging, J. J. Traisnel, A. R. Tannenbaum, Georgia Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-04] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am 10:50 am: Multiple wavelet coherence analysis, S. C. Olhede, G. Metikas, Imperial College London (United Kingdom) . . . . . . . . [6064A-05] 11:10 am: New class of interpolation methods based on discretized lie group transforms, A. Zaratsyan, J. Patera, Univ. de Montrйal (Canada); H. Zhu, York Univ. (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-06] 11:30 am: Optimization procedures for the estimation of phase portrait parameters in orientation fields, F. J. Ayres, R. M. Rangayyan, Univ. of Calgary (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-07] 11:50 am: Optimized gradient filters for hexagonal matrices, T. Shima, S. Saito, Tokyo Institute of Technology (Japan); M. Nakajima, Tokyo Institute of Technology (Japan) and National Institute of Informatics (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-08] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 2:00 pm SESSION 2 Conv. Ctr. Room C2 . . . . . . . . . . . . . Mon. 2:00 to 3:00 pm Efficient Algorithms 2:00 pm: Super-fast Fourier transform, S. S. Agaian, O. Caglayan, The Univ. of Texas at San Antonio . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-09] 2:20 pm: A high-speed rotation method for binary document images based on coordinate operation of run data, Y. Shima, H. Ohya, Meisei Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-11] 2:40 pm: A hardware implementation of the discrete Pascal transform for image processing, T. J. Goodman, M. F. Aburdene, Bucknell Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-12] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm
SESSION 3 Conv. Ctr. Room C2 . . . . . . . . . . . . . Mon. 3:30 to 4:50 pm Image Processing Systems 3:30 pm: Using clustering for document reconstruction, A. Ukovich, A. Zacchigna, G. Ramponi, G. Schoier, Univ. Degli Studi di Trieste (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-13] 3:50 pm: Automatic detection and tracking of reappearing targets in forward-looking infrared imagery, A. Bal, M. S. Alam, Univ. of South Alabama . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-14] 4:10 pm: Robust human motion detection via fuzzy set based image understanding, Q. Li, J. You, The Hong Kong Polytechnic Univ. (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-16] 4:30 pm: k-max: segmentation based on selection of max-tree deep nodes, A. G. Silva, S. C. Felipussi, G. L. F. Cassol, Univ. do Estado de Santa Catarina (Brazil); R. de Alencar Lotufo, Univ. Estadual de Campinas (Brazil) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-17] Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 4 Conv. Ctr. Room C2 . . . . . . . . . . . . Tues. 9:40 to 11:50 am Image Processing Methods 9:40 am: Shape-adaptive DCT for denoising and image reconstruction, A. Foi, K. Dabov, V. Katkovnik, K. O. Egiazarian, Tampere Univ. of Technology (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-18] 10:00 am: Anisotropic filtering with nonlinear structure tensors, C. A. Castaсo Moraga, J. Ruiz-Alzola, Univ. de Las Palmas de Gran Canaria (Spain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-19] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am 10:50 am: A modified wavelet transformation based method of linear object extraction, T. Chen, Univ. of South Carolina . . . . . . . . [6064A-20] 11:10 am: 2D approaches to 3D watermarking: state of the art and perspectives, M. P. Mitrea, S. A. Duta, F. J. Preteux, Institut National des Tйlйcommunications (France) . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-21] 11:30 am: Region-based perceptual grouping: a cooperative approach based on Dempster-Shafer theory, N. Zlatoff, B. Tellez, A. M. Baskurt, Univ. Claude Bernard Lyon 1 (France) . . . . . . . . . . . . . . . . . . . [6064A-22] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:50 am to 1:40 pm
32
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6064A · Conv. Ctr. Room C2
SESSION 5 Conv. Ctr. Room C2 . . . . . . . . . . . . . Tues. 1:40 to 3:00 pm Biomedical Image Processing 1:40 pm: Study of muscular deformation based on surface slope estimation, M. Carli, M. Goffredo, M. Schmid, A. Neri, Univ. degli Studi di Roma Tre (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-23] 2:00 pm: An automated diagnosis approach based on histopathological images, M. C. d'Ornellas, C. C. Danesi, J. A. T. Borges da Costa, Univ. Federal de Santa Maria (Brazil) . . . . . . . . . . . . [6064A-24] 2:20 pm: Variational segmentation of x-ray image with overlapped objects, G. Yu, Nuctech Co. Ltd. (China); L. Zhang, J. Zhang, Y. Xing, H. Gao, Tsinghua Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-25] 2:40 pm: Image segmentation for automated dental identification, E. Haj Said, D. E. M. Nassar, H. H. Ammar, West Virgina Univ. . . [6064A-26] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm SESSION 6 Conv. Ctr. Room C2 . . . . . . . . . . . . . Tues. 3:30 to 5:30 pm Algorithms and Systems 3:30 pm: Deblending of the UV photometry in GALEX deep surveys using optical priors in the visible wavelengths, A. Llebaria, Lab. d'Astrophysique de Marseille (France); M. Guillaume, D. Aymeric, Ecole Generaliste d'Ingenieurs de Marseille (France); B. Milliard, S. Arnauts, Lab. d'Astrophysique de Marseille (France) . . . . . . . . . . . . . . . . . . . [6064A-27] 3:50 pm: Comparative study of logarithmic enhancement algorithms with performance measure, E. J. Wharton, Tufts Univ.; S. S. Agaian, The Univ. of Texas at San Antonio; K. A. Panetta, Tufts Univ. . . . . . [6064A-28] 4:10 pm: MMW video sequence denoising and enhancement in concealed weapons detection applications, X. Wei, H. Chen, The Univ. of Texas at Arlington; P. K. Varshney, Syracuse Univ. . . . . . . . . [6064A-29] 4:30 pm: Image denoising with block-matching and 3D filtering, K. Dabov, A. Foi, V. Katkovnik, K. O. Egiazarian, Tampere Univ. of Technology (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-30] 4:50 pm: An algorithm for the enhancement of images of large dynamic range, F. Hassan, J. E. Carletta, Univ. of Akron . . . . [6064A-31] 5:10 pm: Nonlinear image enhancement to improve face detection in complex lighting environment, L. Tao, M. Seow, V. K. Asari, Old Dominion Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-32]
Posters and Demonstrations-Tuesday Demonstrations. . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Noise image enhancement on Hцlder function spaces, M. Lee, National Taiwan Ocean Univ. (Taiwan) . . . . . . . . . . . . . . . . . [6064A-33] Ad hoc segmentation pipeline for microarray image analysis, S. Battiato, G. Di Blasi, G. M. Farinella, G. Gallo, G. C. Guarnera, Univ. di Catania (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-34] Color transient improvement via range detection, G. S. Shin, M. G. Kang, Yonsei Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . [6064A-35] For fast classification and segmentation of high-resolution images of multiple and complicated colonies, W. Wang, Chongqing Univ. of Posts and Telecommunications (China) . . . . . . . . . . . . . . . . [6064A-36] Spatially adaptive multiresolution multispectral image fusion based on Bayesian approach, J. H. Park, M. G. Kang, Yonsei Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-37] Segmentation of microspheres in ultrahigh density microspherebased assays, A. Mathur, D. M. Kelso, Northwestern Univ. [6064A-38] Discrete Gould transform and its applications, M. F. Aburdene, H. M. Le, Bucknell Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-39] The application of image filters combined with the nonlinear regression analysis on optical autofocusing, M. Lee, W. Hsu, National Kaohsiung Normal Univ. (Taiwan); T. Lin, National Taiwan Univ. (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-40] Crease enhancement and the segmentation of topographic images, H. C. Morris, San Josй State Univ. . . . . . . . . . . . . . . . . . . . . [6064A-41] A multiscale approach to contour detection and texture suppression, G. Papari, Rijksuniv. Groningen (Netherlands); P. Campisi, Univ. degli Studi Roma Tre (Italy); N. Petkov, Rijksuniv. Groningen (Netherlands); A. Neri, Univ. degli Studi Roma Tre (Italy) . . . [6064A-43] A heuristic approach for the extraction of region and boundary of mammalian cells in bio-electric images, L. Tao, V. K. Asari, Old Dominion Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-44] Lip segmentation and tracking for facial palsy, M. Park, J. Seo, K. S. Park, Seoul National Univ. (South Korea) . . . . . . . . . . . . . . . [6064A-45] A multithreshold algorithm based on intensity distribution, D. Chen, M. Sarhadi, Brunel Univ. (United Kingdom) . . . . . . . . . . . . . [6064A-46] An efficient multiresolution GA approach to dental image alignment, D. E. M. Nassar, M. Ogirala, D. A. Adjeroh, H. H. Ammar, West Virginia Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064A-47] Phase unwrapping by means of finite differences, L. I. Olivos-Pйrez, E. de la Rosa Miranda, L. R. Berriel-Valdos, R. Ramos-Lуpez, Instituto Nacional de Astrofнsica, Уptica y Electrуnica (Mexico) . . . . [6064A-48]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
33
Conference 6064B · Conv. Ctr. Room C2
Tuesday-Wednesday 17-18 January 2006 · Part of Proceedings of SPIE Vol. 6064 Image Processing: Algorithms and Systems, Neural Networks and Machine Learning Applications of Neural Networks and Machine Learning in Image Processing X Conference Chairs: Nasser M. Nasrabadi, Army Research Lab.; Syed A. Rizvi, CUNY/College of Staten Island Program Committee: Pierre Baldi, California Institute of Technology; Yoshua Bengio, Univ. de Montrйal (Canada); Terry M. Caelli, Curtin Univ. of Technology (Australia); Rama Chellappa, Univ. of Maryland/College Park; Chang Y. Choo, San Josй State Univ.; Sandor Z. Der, Aerospace Corp.; Edward R. Dougherty, Texas A&M Univ.; Kunihiko Fukushima, Tokyo Univ. of Technology (Japan); Erol Gelenbe, Imperial College London (United Kingdom); David H. Haussler, Univ. of California/Santa Cruz; Nicolaos B. Karayiannis, Univ. of Houston; Bart Kosko, Univ. of Southern California; Sun-Yuan Kung, Princeton Univ.; Richard P. Lippmann, MIT Lincoln Lab.; Erkki Oja, Helsinki Univ. of Technology (Finland); Sankar K. Pal, Indian Statistical Institute (India); Tomaso A. Poggio, MIT Artificial Intelligence Lab.; Christoph von der Malsburg, Univ. of Southern California; Jacek M. Zurada, Univ. of Louisville
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Key-text spotting in documentary videos using Adaboost, M. Lalonde, L. Gagnon, CRIM (Canada) . . . . . . . . . . . . . . . . . . [6064B-65] Research on classifying performance of SVMs with basic kernel in HCCR, L. Sun, YanTai Univ. (China) . . . . . . . . . . . . . . . . . . . [6064B-67] Face recognition based on HMM in compressed domain, H. Wang, G. Feng, Sun Yat-sen Univ. (China) . . . . . . . . . . . . . . . . . . . [6064B-68] Application of ANN and DT on medium resolution ASTER image to model gully network in Southern Italy, A. Ghaffari, P. M. Mather, G. Priestnall, M. L. Clarke, The Univ. of Nottingham (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-69]
Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 7 Conv. Ctr. Room C2 . . . . . . . . . . . . Wed. 9:30 to 10:30 am Kernel-based Learning for Detection and Shape Analysis Chair: Heesung Kwon, Army Research Lab. 9:30 am: Nonlinear shape prior from Kernel space for geometric active contours, S. Dambreville, Y. Rathi, A. R. Tannenbaum, Georgia Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-49] 9:50 am: Kernel subspace matched target detectors, H. Kwon, N. M. Nasrabadi, Army Research Lab. . . . . . . . . . . . . . . . . . . . . . . . . [6064B-50] 10:10 am: Statistical shape analysis using kernel PCA, Y. Rathi, S. Dambreville, A. R. Tannenbaum, Georgia Institute of Technology[6064B-51] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am SESSION 8 Conv. Ctr. Room C2 . . . . . . . . . . . Wed. 11:00 to 11:20 am Fuzzy Clustering Chair: Heesung Kwon, Army Research Lab. 11:00 am: Segmentation and enhancement of digital copies using a new fuzzy clustering method, M. N. Ahmed, B. E. Cooper, Lexmark International, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-52]
34
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6064B · Conv. Ctr. Room C2 SESSION 9 Conv. Ctr. Room C2 . . . . . . . . . . . Wed. 11:20 to 11:40 am Independent Component Analysis, Adaboost for Recognition Chair: Syed A. Rizvi, College of Staten Island/CUNY 11:20 am: 2D/3D facial feature extraction, B. Sankur, L. Akarun, H. Cinar, A. Ali Salah, Bogaziзi Univ. (Turkey) . . . . . . . . . . . . . . . . . . . . . [6064B-54] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:40 am to 1:20 pm SESSION 10 Conv. Ctr. Room C2 . . . . . . . . . . . . . Wed. 1:20 to 3:50 pm Neural Networks Applications for Manifold Learning, Recognition, Color Perception, and Compression Chair: Syed A. Rizvi, College of Staten Island/CUNY 1:20 pm: Noniterative neural network learning of an N-dimension curve representing the dynamic history of a time varying pattern, C. J. Hu, Southern Illinois Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-56] 1:40 pm: Manifold of color perception: color constancy using a nonlinear line attractor, V. K. Asari, M. Seow, Old Dominion Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-57] 2:00 pm: A novel neural net application for image compression, H. S. Soliman, M. Omari, New Mexico Institute of Mining and Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-58] 2:20 pm: Toward content-based object recognition with image primitives, G. Wang, J. M. Kinser, George Mason Univ. . . . . . [6064B-59] 2:40 pm: Translation invariance in a network of oscillatory units, A. R. Rao, G. A. Cecchi, C. Peck, J. Kozloski, IBM Corp. . . . . . . . . . [6064B-60] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm 3:30 pm: Efficient learning and recognition using segmented analytical data of an edge-detected 2D image, C. J. Hu, Southern Illinois Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-61] SESSION 11 Conv. Ctr. Room C2 . . . . . . . . . . . . . Wed. 3:50 to 4:50 pm Support Vector Machine and Neural Networks for Face Recognition, Detection, and Classification Chair: Nasser M. Nasrabadi, Army Research Lab. 3:50 pm: Support vector machine as digital image watermark detector, P. H. H. Then, Swinburne Univ. of Technology (Malaysia); Y. C. Wang, Univ. Malaysia Sarawak (Malaysia) . . . . . . . . . . . . . . . . . . . . . . . . . . [6064B-62] 4:10 pm: Neural networks approach to high vertical resolution atmospheric temperature profile retrieval from spaceborne high spectral resolution infrared sounder measurements, D. Jiang, C. Dong, Hunan Meterological Bureau (China) . . . . . . . . . . . . . . . . . . . . [6064B-63] 4:30 pm: Probabilistic multiresolution human classification, H. Ran, J. Tu, Wuhan Univ. of Technology (China) . . . . . . . . . . . . . . . . . . [6064B-64]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
35
Conference 6065 · Conv. Ctr. Room A4
Monday-Wednesday 16-18 January 2006 · Proceedings of SPIE Vol. 6065 Computational Imaging IV
Conference Chairs: Charles A. Bouman, Purdue Univ.; Eric L. Miller, Northeastern Univ.; Ilya Pollak, Purdue Univ. Program Committee: Thomas S. Denney, Jr., Auburn Univ.; Peter C. Doerschuk, Purdue Univ.; Maya R. Gupta, Univ. of Washington; Peyman Milanfar, Univ. of California/Santa Cruz; Joseph A. O'Sullivan, Washington Univ. in St. Louis; Zygmunt Pizlo, Purdue Univ.; Stanley J. Reeves, Auburn Univ.; Yongyi Yang, Illinois Institute of Technology; Yibin Zheng, Univ. of Virginia
Monday 16 January SESSION 1 Conv. Ctr. Room A4 . . . . . . . . . . . . . Mon. 8:30 to 9:00 am Keynote Presentation I Chair: Charles A. Bouman, Purdue Univ. Keynote 8:30 am: Keynote (Invited Paper), M. V. de Hoop, Purdue Univ. [6065-01] SESSION 2 Conv. Ctr. Room A4 . . . . . . . . . . . . Mon. 9:00 to 11:10 am Hierarchical and Graph-based Image Analysis Chair: Peyman Milanfar, Univ. of California/Santa Cruz 9:00 am: Modeling hierarchical structure of images with stochastic grammars, W. Wang, T. Wong, I. Pollak, C. A. Bouman, M. P. Harper, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-02] 9:20 am: Multiresolution analysis of digital images using the continuous extension of discrete group transforms, M. Germain, J. Patera, A. Zaratsyan, Univ. de Montrйal (Canada) . . . . . . . . . . . . [6065-03] 9:40 am: Modeling multiscale differential pixel statistics with applications, D. Odom, P. Milanfar, Univ. of California/Santa Cruz[6065-04] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am 10:30 am: Graph-based 3D object classification, S. Baloch, A. H. Krim, North Carolina State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-05] 10:50 am: Compression via optimal basis selection in large treestructured dictionaries, Y. Huang, Purdue Univ. . . . . . . . . . . . . [6065-06] SESSION 3 Conv. Ctr. Room A4 . . . . . . . . Mon. 11:10 am to 12:10 pm Reconstruction from Sparse Data Chair: Ilya Pollak, Purdue Univ. 11:10 am: Compressed sensing in noisy imaging environments, J. Haupt, R. Castro, R. D. Nowak, Univ. of Wisconsin/Madison . . . [6065-07] 11:30 am: Stable signal recovery from incomplete and inaccurate observations, J. K. Romberg, California Institute of Technology [6065-08] 11:50 am: A compressed sensing camera: new theory and an implementation using digital micromirrors, D. Takhar, M. B. Wakin, M. Duarte, D. Baron, K. F. Kelly, R. G. Baraniuk, Rice Univ. . . . . . . . [6065-09] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:40 pm
SESSION 4 Conv. Ctr. Room A4 . . . . . . . . . . . . . Mon. 1:40 to 3:00 pm Microscopy Chair: Peter C. Doerschuk, Purdue Univ. 1:40 pm: A fast algorithm for 3D reconstruction from unoriented projections and cryo-electron microscopy of viruses, J. Lee, Y. Zheng, P. C. Doerschuk, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-10] 2:00 pm: Spatially adaptive 3D inverse for optical sectioning, D. V. Paliy, V. Katkovnik, K. O. Egiazarian, Tampere Univ. of Technology (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-11] 2:20 pm: On soft clipping of Zernike moments for deblurring and enhancement of optical point spread functions, N. Becherer, J. Hesser, Univ. Mannheim (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-12] 2:40 pm: Adaptive sampling for atomic force microscopy with system level motion constraints, H. Cheng, G. T. C. Chiu, Purdue Univ. [6065-13] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm SESSION 5 Conv. Ctr. Room A4 . . . . . . . . . . . . . Mon. 3:30 to 5:10 pm Inverse Problems Chair: Thomas S. Denney, Jr., Auburn Univ. 3:30 pm: Bayesian image reconstruction from Fourier-domain samples using prior edge information: convergence and parameter sensitivity, T. S. Denney, Jr., S. J. Reeves, Auburn Univ. . . . . . . . . . . . . . . . . [6065-14] 3:50 pm: Thin digital imaging systems using focal plane coding, A. D. Portnoy, J. P. Guo, N. P. Pitsianis, D. J. Brady, Duke Univ.; M. A. Fiddy, Univ. of North Carolina/Charlotte; M. R. Feldman, R. D. Te Kolste, Digital Optics Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-15] 4:10 pm: 3D reconstructions from spherically averaged Fourier transform magnitude and solution x-ray scattering experiments, Y. Hwang, P. C. Doerschuk, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . [6065-16] 4:30 pm: Computed spectroscopy using segmented apertures, R. T. Hoctor, F. W. Wheeler, GE Global Research; E. B. Barrett, Lockheed Martin Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-17] 4:50 pm: Preconditioned conjugate gradient without linesearch: a comparison with the half-quadratic approach for edge-preserving image restoration, C. Labat, J. Idier, Institute of Research in Communications and Cybernetics of Nantes (France) . . . . . . . . . [6065-18]
36
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6065 · Conv. Ctr. Room A4
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 6 Conv. Ctr. Room A4 . . . . . . . . . . . . Tues. 9:30 to 10:00 am Keynote Presentation II Chair: Eric L. Miller, Northeastern Univ. 9:30 am: Computational methods for image restoration, image segmentation, and texture modeling (Invited Paper), G. Chung, T. M. Le, L. H. Lieu, N. Tanushev, L. Vese, Univ. of California/Los Angeles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-19] SESSION 7 Conv. Ctr. Room A4 . . . . . . . . Tues. 10:00 am to 12:10 pm Image and Video Analysis Chair: Mireille Boutin, Purdue Univ. 10:00 am: An adaptive model for restoration of optically distorted video frames, D. Li, Georgia Institute of Technology; M. J. T. Smith, Purdue Univ.; R. M. Mersereau, Georgia Institute of Technology . [6065-20] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am 10:50 am: Resource-driven content adaptation, Y. Lu, D. S. Ebert, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-21] 11:10 am: Algebraic methods for structure from motion, M. Boutin, J. Zhang, D. G. Aliaga, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . [6065-22] 11:30 am: A maximum entropy kernel density estimator with applications to function interpolation and texture segmentation, N. Balakrishnan, D. Schonfeld, Univ. of Illinois at Chicago . . . . . . . . [6065-23] 11:50 am: Multiple watermarking: a vector space projection approach, O. Altun, G. Sharma, M. Bocko, Univ. of Rochester . . . . . . . . . . . [6065-24] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:40 pm SESSION 8 Conv. Ctr. Room A4 . . . . . . . . . . . . . Tues. 1:40 to 3:00 pm Biomedical Imaging Chair: Miles N. Wernick, Illinois Institute of Technology 1:40 pm: Spherical harmonics for shape-based inverse problems, as applied to electrical impedance tomography, S. Babaeizadeh, D. H. Brooks, Northeastern Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-25] 2:00 pm: 3D nonlinear multigrid algorithm for direct reconstruction of chromophore concentrations in diffuse optical tomography, J. C. Ye, Korea Advanced Institute of Science and Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-26] 2:20 pm: Adaptation of fast marching methods to subcellular modeling, A. Chikando, J. M. Kinser, George Mason Univ. . . . . . [6065-27] 2:40 pm: Machine learning of human responses to images, M. N. Wernick, Y. Yang, J. G. Brankov, L. Wei, Illinois Institute of Technology; I. M. El-Naqa, Washington Univ. in St. Louis; N. P. Galatsanos, Univ. of Ioannina (Greece) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-28] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm
SESSION 9 Conv. Ctr. Room A4 . . . . . . . . . . . . . Tues. 3:30 to 5:50 pm Tomography Chair: Joseph A. O'Sullivan, Washington Univ. in St. Louis 3:30 pm: Image reconstruction algorithms for a novel PET system with a half-ring insert, D. Pal, J. A. O'Sullivan, H. Wu, M. Janecek, Y. C. Tai, Washington Univ. in St. Louis . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-29] 3:50 pm: Improved sampling of parallel projection in cylindrical PET scanners, B. Farsaii, SUNY/Univ. at Buffalo . . . . . . . . . . . . . . . . [6065-30] 4:10 pm: A Bayesian approach to tomography of multiply scattered beams, Z. H. Levine, National Institute of Standards and Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-31] 4:30 pm: Progress in multiple-image radiography, M. N. Wernick, J. G. Brankov, Y. Yang, G. Khelashvili, Illinois Institute of Technology; D. Chapman, Univ. of Saskatchewan (Canada); I. Mondal, B. Marquet, Illinois Institute of Technology; Z. Zhong, Brookhaven National Lab. . . . [6065-32] 4:50 pm: A recursive filter for noise reduction in tomographic imaging, J. Thibault, GE Medical Systems; C. A. Bouman, Purdue Univ.; J. Hsieh, GE Medical Systems; K. D. Sauer, Univ. of Notre Dame . . . . . . . [6065-33] 5:10 pm: A branch-less distance driven projection and backprojection algorithm, S. Basu, General Electric Co.; B. De Man, GE Global Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-34] 5:30 pm: Cupping artifacts analysis and correction for a FPD-based cone-beam CT, L. Zhang, H. Gao, Z. Chen, S. Li, Y. Xing, Tsinghua Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-35] Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. A block-iterative deterministic annealing algorithm for Bayesian tomographic reconstruction, S. Lee, Paichai Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-44] Deinterlacing in spatial and temporal domain, I. Kim, C. Lee, Yonsei Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-46] Cosine transform generalized to lie groups SU(2)xSU(2), O(5) and SU(2)xSU(2)xSU(2): application to digital image processing, M. Germain, J. Patera, Univ. de Montrйal (Canada) . . . . . . . . . . . [6065-47] A prioritized and adaptive approach to volumetric seeded region growing using texture descriptors, N. J. Backman, Whitworth College; B. W. Whitney, Northern Kentucky Univ.; J. D. Furst, D. S. Raicu, DePaul Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-48] A fast MAP-based superresolution algorithm for general motion, M. Tanaka, M. Okutomi, Tokyo Insititute of Technology (Japan) . . [6065-49] Image deblurring by the combined use of a superresolution technique and inverse filtering, Y. Yamada, K. Nakamae, Osaka Univ. (Japan); H. Fujioka, Fukui Univ. of Technology (Japan) . . . . . . [6065-50]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
37
Conference 6065 · Conv. Ctr. Room A4 Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 10 Conv. Ctr. Room A4 . . . . . . . . . . . . Wed. 9:30 to 10:30 am Color Chair: Stanley J. Reeves, Auburn Univ. 9:30 am: Estimation of color filter array data from JPEG images for improved demosaicing, W. Feng, S. J. Reeves, Auburn Univ. . . [6065-36] 9:50 am: Separation of irradiance and reflectance from observed color images by logarithmical nonlinear diffusion process, T. Saito, H. Takahashi, T. Komatsu, Kanagawa Univ. (Japan) . . . . . . . . . . . . . [6065-37] 10:10 am: Novel scanner characterization method for color measurement and diagnostics applications, B. Lee, Thomson Corporate Research; R. Bala, Xerox Corp.; G. Sharma, Univ. of Rochester . [6065-38] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am SESSION 11 Conv. Ctr. Room A4 . . . . . . . . Wed. 11:00 am to 12:20 pm Image Modeling and Analysis Chair: Zygmunt Pizlo, Purdue Univ. 11:00 am: Elastic surface registration by parameterization optimization in spectral space, F. G. Vadakkumpadan, Y. Tong, Y. Sun, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-39] 11:20 am: Mosaicking of astronomical images with MOPEX, D. Makovoz, I. Khan, F. J. Masci, California Institute of Technology . [6065-40] 11:40 am: Image processing using parallel GPU units, K. A. Bjorke, NVIDIA Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6065-41] 12:00 pm: Partial shape similarity of contours is needed for object recognition, Z. Pizlo, Purdue Univ.; L. J. Latecki, Temple Univ. . [6065-42]
38
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6066 · Conv. Ctr. Room C4
Tuesday-Wednesday 17-18 January 2006 · Proceedings of SPIE Vol. 6066 Vision Geometry XIV
Conference Chairs: Longin Jan Latecki, Temple Univ.; David M. Mount, Univ. of Maryland/College Park; Angela Y. Wu, American Univ. Program Committee: Gady Agam, Illinois Institute of Technology; Gilles Bertrand, Groupe ESIEE (France); Atsushi Imiya, Chiba Univ. (Japan); Jack Koplowitz, Clarkson Univ.; Nathan S. Netanyahu, Bar Ilan Univ. (Israel); Peter Veelaert, Hogeschool Gent (Belgium); Xiaodong Wu, Univ. of Iowa
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 1 Conv. Ctr. Room C4 . . . . . . . . . Tues. 9:30 am to 12:00 pm Shape and Object Recognition I Chair: Longin Jan Latecki, Temple Univ. 9:30 am: A deformable model with topology analysis and adaptive clustering for boundary detection, M. Allili, Bishop's Univ. (Canada); B. Yang, Univ. de Sherbrooke (Canada); L. Bentabet, Bishop's Univ. (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-01] 9:55 am: Refining road map using active shape model from aerial images, G. Koutaki, K. Uchimura, Z. Hu, Kumamoto Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-02] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:45 am 10:45 am: Quantification of line-mura defect level based on multiple characterizing features, N. K. Park, K. N. Choi, S. I. Yoo, Seoul National Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-03] 11:10 am: Model-based shape classification using shapetransformation-invariant descriptors, S. C. Lee, Univ. of Oklahoma; Y. Wang, E. T. Lee, Univ. of Oklahoma Health Sciences Ctr. . . . . . . [6066-04] 11:35 am: Refinement of axial shape description, A. N. Skourikhine, Los Alamos National Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-05] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:20 pm SESSION 2 Conv. Ctr. Room C4 . . . . . . . . . . . . . Tues. 1:20 to 2:00 pm Shape and Object Recognition II Chair: Longin Jan Latecki, Temple Univ. 1:20 pm: Geometry of human vision (Invited Paper, Presentation Only), Z. Pizlo, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-06]
SESSION 3 Conv. Ctr. Room C4 . . . . . . . . . . . . . Tues. 2:00 to 3:15 pm 3D Geometry Chair: Kun Lee, Handong Global Univ. (South Korea) 2:00 pm: Fitting polygonal regions for matching 3D polyhedra, L. Mukherjee, V. Singh, J. Xu, R. Berezney, SUNY/Univ. at Buffalo . [6066-07] 2:25 pm: Hierarchical two view line segment matching using wavelet transform, F. Mai, Y. Hung, W. Sze, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-09] 2:50 pm: A vision-based approach to extracting the tilt angle and altitude of a PTZ camera, I. Chen, S. Wang, National Chiao Tung Univ. (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-10] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:15 to 3:45 pm SESSION 4 Conv. Ctr. Room C4 . . . . . . . . . . . . . Tues. 3:45 to 4:35 pm Aspects of Vision Geometry Chair: David M. Mount, Univ. of Maryland/College Park 3:45 pm: Perspex machine V: compilation of C programs, M. P. Spanner, J. A. D. W. Anderson, The Univ. of Reading (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-12] 4:10 pm: Automatic and robust classification of independent motions in video sequences, X. An, Zhejiang Univ. (China) . . . . . . . . . . . [6066-13] Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. A three-dimensional shape measurement method: structure light space-time stereo, X. Li, Shanghai Univ. (China) . . . . . . . . . . [6066-28] Multiview image calibration and rectification for and effective 3D display, K. Bae, 3R Inc. (South Korea); H. Kang, E. Kim, Kwangwoon Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-29] Perspex machine VI: a graphical user interface to the Perspex machine, C. J. Kershaw, J. A. D. W. Anderson, The Univ. of Reading (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-30] Perspex machine VII: the universal Perspex machine, J. A. D. W. Anderson, The Univ. of Reading (United Kingdom) . . . . . . . . . [6066-31]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
39
Conference 6066 · Conv. Ctr. Room C4
Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 5 Conv. Ctr. Room C4 . . . . . . . . . . . . Wed. 9:30 to 11:35 am Digital Geometry and Topology Chair: Peter F. Stiller, Texas A&M Univ. 9:30 am: Discrete circles: an arithmetical approach with non-constant thickness, C. Fiorio, D. Jamet, J. Toutant, Univ. Montpellier II (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-14] 9:55 am: Estimating the analog perimeter of a pre-digitized shape, S. C. Lee, Univ. of Oklahoma; Y. Wang, E. T. Lee, Univ. of Oklahoma Health Sciences Ctr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-15] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:45 am 10:45 am: Three-dimensional fast exact Euclidean distance (3D-FEED) maps, T. E. Schouten, H. C. Kuppens, Radboud Univ. Nijmegen (Netherlands); E. L. van den Broek, Vrije Univ. Amsterdam (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-17] 11:10 am: Estimating the surface area and volume of a general 3D shape, S. C. Lee, Univ. of Oklahoma; Y. Wang, E. T. Lee, Univ. of Oklahoma Health Sciences Ctr. . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-18] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:35 am to 1:20 pm SESSION 6 Conv. Ctr. Room C4 . . . . . . . . . . . . . Wed. 1:20 to 3:25 pm Image Matching and Registration Chair: James G. Anderson, The Univ. of Reading (United Kingdom) 1:20 pm: Dynamic RANSAC, W. Sze, A. W. Tang, Y. Hung, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-19] 1:45 pm: Singular value decomposition based scale invariant image matching, W. Sze, A. W. Tang, Y. Hung, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-20] 2:10 pm: Image matching using algebraic topology, S. Derdar, Univ. de Sherbrooke (Canada); A. Madjid, Bishop's Univ. (Canada); D. Ziou, Univ. de Sherbrooke (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-21] 2:35 pm: Robustness and statistical analysis of object/image metrics, P. F. Stiller, Texas A&M Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-22] 3:00 pm: GSIFT: Geometric Scale Invariant Feature Transform for Terrain Data, Y. Xiao, S. K. Lodha, Univ. of California/Santa Cruz [6066-23] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:25 to 3:50 pm
SESSION 7 Conv. Ctr. Room C4 . . . . . . . . . . . . . Wed. 3:50 to 5:30 pm Surface Reconstruction and Visualization Chair: Samuel C. Lee, Univ. of Oklahoma 3:50 pm: Reconstruction of auadratic curves in 3D using two or more perspective views: simulation studies, S. Kumar, N. Sukavanam, R. Balasubramanian, Indian Institute of Technology Roorkee (India) [6066-24] 4:15 pm: Visualization of volumetric scattered data by using weighted alpha shapes, J. Paik, K. Lee, Handong Global Univ. (South Korea); O. Gwun, Chonbuk National Univ. (South Korea) . . . . . . . . . . . . . . . [6066-25] 4:40 pm: POSS: efficient nonlinear optimization for parameterization methods, F. G. Vadakkumpadan, Y. Tong, Y. Sun, Purdue Univ. . [6066-26] 5:05 pm: Piecewise compression of large mesh, A. Qin, Zhejiang Univ. (China) and Shanxi Univ. (China); J. Shi, Z. Liu, M. Huang, Zhejiang Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6066-27]
40
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6067 · Conv. Ctr. Room C3
Wednesday-Thursday 18-19 January 2006 · Proceedings of SPIE Vol. 6067 Document Recognition and Retrieval XIII
Conference Chairs: Kazem Taghva, Univ. of Nevada/Las Vegas; Xiaofan Lin, Hewlett-Packard Labs. Program Committee: Tim L. Andersen, Boise State Univ.; Apostolos Antonacopoulos, Univ. of Salford (United Kingdom); Elisa H. Barney Smith, Boise State Univ.; Brian D. Davison, Lehigh Univ.; Xiaoqing Ding, Tsinghua Univ. (China); David S. Doermann, Univ. of Maryland/ College Park; Jianying Hu, IBM Thomas J. Watson Research Ctr.; Matthew F. Hurst, Intelliseek, Inc.; Hisashi Ikeda, Hitachi, Ltd. (Japan); Tapas Kanungo, IBM Almaden Research Ctr.; Daniel P. Lopresti, Lehigh Univ.; Thomas A. Nartker, Univ. of Nevada/Las Vegas; Sargur N. Srihari, SUNY/Univ. at Buffalo; George R. Thoma, National Library of Medicine; Berrin A. Yanikoglu, Sabanci Univ. (Turkey)
Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 1 Conv. Ctr. Room C3 . . . . . . . . . . . . Wed. 9:30 to 10:30 am Handwriting Recognition 9:30 am: Combining one- and two-dimensional signal recognition approaches to off-line signature verification, S. N. Srihari, SUNY/Univ. at Buffalo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-01] 9:50 am: Spotting words in handwritten Arabic documents, S. N. Srihari, H. Srinivasan, P. Babu, C. Bhole, SUNY/Univ. at Buffalo . [6067-02] 10:10 am: HCCR by contour-based elastic mesh fuzzy feature, L. Sun, YanTai Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-03] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am SESSION 2 Conv. Ctr. Room C3 . . . . . . . . . . . Wed. 11:00 to 11:40 am Invited Paper I 11:00 am: Human language technology research at DARPA (Invited Paper, Presentation Only), J. Olive, Defense Advanced Research Projects Agency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-04] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:40 am to 1:40 pm SESSION 3 Conv. Ctr. Room C3 . . . . . . . . . . . . . Wed. 1:40 to 3:00 pm Optical Character Recognition 1:40 pm: Partitioning of the degradation space for OCR training, E. H. Barney Smith, T. L. Andersen, Boise State Univ. . . . . . . . . . . . . . [6067-05] 2:00 pm: Match graph generation for symbolic indirect correlation, D. P. Lopresti, Lehigh Univ.; G. Nagy, A. Joshi, Rensselaer Polytechnic Institute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-06] 2:20 pm: Toward quantifying the amount of style in a dataset, X. Zhang, S. Andra, Rensselaer Polytechnic Institute . . . . . . . . . . . . . . . . . . [6067-07] 2:40 pm: Robust feature extraction for character recognition based on binary image, L. Wang, L. Zhang, Y. Xing, Z. Wang, Nuctech Co. Ltd. (China); H. Gao, Tsinghua Univ. (China) . . . . . . . . . . . . . . . . . . . . [6067-08] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm
SESSION 4 Conv. Ctr. Room C3 . . . . . . . . . . . . . Wed. 3:30 to 5:10 pm Image Processing 3:30 pm: DOCLIB: a software library for document processing, S. R. Jaeger, G. Zhu, D. S. Doermann, Univ. of Maryland/College Park [6067-09] 3:50 pm: Address block features for image-based automated mail orientation, M. S. Khan, SRI International and Univ. of California; H. B. Aradhye, SRI International . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-10] 4:10 pm: A robust stamp detection framework on degraded documents, G. Zhu, S. R. Jaeger, D. S. Doermann, Univ. of Maryland/ College Park . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-11] 4:30 pm: Adaptive pre-OCR cleanup of grayscale document images, I. Zavorin, E. Borovikov, M. I. Turner, CACI International Inc.; L. Hernandez, Army Research Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-12] 4:50 pm: JBIG2 text image compression based on OCR, J. Shang, C. Liu, X. Ding, Tsinghua Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . [6067-13] Thursday 19 January SESSION 5 Conv. Ctr. Room C3 . . . . . . . . . . . Thurs. 8:40 to 10:20 am Emerging Applications 8:40 am: Active document versioning: from layout understanding to adjustment, X. Lin, H. Chao, G. Nelson, E. Durante, Hewlett-Packard Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-14] 9:00 am: Graphic design principles for automated document segmentation and understanding, F. Vega, H. J. Santos-Villalobos, Univ. de Puerto Rico Mayagьez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-15] 9:20 am: A new document authentication method by embedding deformation characters, X. Wang, X. Ding, H. Liu, C. Liu, Tsinghua Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-16] 9:40 am: CAPTCHA challenge strings: problems and improvements, J. L. Bentley, C. L. Mallows, Avaya Inc. . . . . . . . . . . . . . . . . . . . . . . [6067-17] 10:00 am: An automatically updateable web publishing solution: taking document sharing and conversion to enterprise level, F. Rahman, BCL Technologies Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-18] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
41
Conference 6067 · Conv. Ctr. Room C3 SESSION 6 Conv. Ctr. Room C3 . . . . . . . . . . Thurs. 10:50 to 11:50 am Document Retrieval 10:50 am: Automatic redaction of private information using relational information extraction, K. Taghva, R. Beckley, J. S. Coombs, J. Borsack, R. Pereda, T. A. Nartker, Univ. of Nevada/Las Vegas . . . . . . . . . . [6067-19] 11:10 am: Document clustering: applications in a collaborative digital library, F. Rahman, BCL Technologies Inc. . . . . . . . . . . . . . . . . . . [6067-20] 11:30 am: Author name recognition in degraded journal images, A. de Bodard de la Jacopiere, L. Likforman, Йcole Nationale Supйrieure des Tйlйcommunications (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-21] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:50 am to 1:40 pm SESSION 7 Conv. Ctr. Room C3 . . . . . . . . . . . . Thurs. 1:40 to 2:20 pm Invited Paper II 1:40 pm: Complex document information processing: prototype, test collection, and evaluation (Invited Paper), G. Agam, Illinois Institute of Technology; S. Argamon, inois Institute of Technology; O. Frieder, D. Grossman, D. Lewis, Illinois Institute of Technology . . . . . . . . . . [6067-23] SESSION 8 Conv. Ctr. Room C3 . . . . . . . . . . . . Thurs. 2:20 to 3:40 pm Learning and Classification 2:20 pm: Comparative evaluation of different classifiers for robust distorted character recognition, B. As-Sadhan, Z. Al Bawab, A. El Seed, M. Noamany, Carnegie Mellon Univ. . . . . . . . . . . . . . . . . . . . . . . . [6067-24] 2:40 pm: Style consistent nearest neighbor classifier, S. Andra, X. Zhang, Rensselaer Polytechnic Institute . . . . . . . . . . . . . . . . . . . [6067-25] 3:00 pm: Optimally combining a cascade of classifiers, K. H. Chellapilla, M. M. Shilman, P. Simard, Microsoft Corp. . . . . . . . . . . . . . . . . . . [6067-26] 3:20 pm: Versatile document image content extraction, H. S. Baird, M. A. Moll, Lehigh Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6067-27]
42
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6068 · Conv. Ctr. Room A6
Wednesday-Thursday 18-19 January 2006 · Proceedings of SPIE Vol. 6068 Sensors, Cameras, and Systems for Scientific/Industrial Applications VIII
Conference Chair: Morley M. Blouke, Ball Aerospace & Technologies Corp. Program Committee: Erik Bodegom, Portland State Univ. and Technische Univ. Delft (Netherlands); Robin M. Dawson, Sarnoff Corp.; Terrence S. Lomheim, The Aerospace Corp.; Gloria G. Putnam, Eastman Kodak Co.; Alice L. Reinheimer, e2v Technologies, Inc.; Nobukazu Teranishi, Matsushita Electric Industrial Co., Ltd. (Japan); Orly Yadid-Pecht, Univ. of Calgary (Canada)
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. An improved method for calculating the MTF of an optical system, A. Walter, S. Lashansky, Israel Air Force (Israel) . . . . . . . . . . . . [6068-30] Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 1 Conv. Ctr. Room A6 . . . . . . . . . . . . Wed. 9:30 to 11:40 am CMOS Devices 9:30 am: In vitro and in vivo on-chip biofluorescence imaging using a CMOS image sensor, D. C. Ng, M. Matsuo, T. Tokuda, K. Kagawa, M. Nunoshita, J. Ohta, Nara Institute of Science and Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-01] 9:50 am: An optical and potential dual-image CMOS sensor for bioscientific applications, T. Tokuda, A. Yamamoto, K. Kagawa, M. Nunoshita, J. Ohta, Nara Institute of Science and Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-02] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am
10:40 am: A CMOS active pixel sensor for retinal stimulation, M. L. Prydderch, M. J. French, Rutherford Appleton Lab. (United Kingdom); K. Mathieson, C. Adams, D. Gunning, J. Laudanski, J. D. Morrison, A. R. Moodie, J. Sinclair, Univ. of Glasgow (United Kingdom) . . . . . . . [6068-03] 11:00 am: Noise analysis of fault tolerant active pixel sensors with and without defects, M. L. La Haye, C. G. Jung, M. H. Izadi, G. H. Chapman, K. S. Karim, Simon Fraser Univ. (Canada) . . . . . . . . . . . . . . . . . . [6068-04] 11:20 am: A model for dark current characterization and simulation, R. L. Baer, Agilent Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-05] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:40 am to 1:30 pm SESSION 2 Conv. Ctr. Room A6 . . . . . . . . . . . . . Wed. 1:30 to 4:20 pm CMOS Devices and Applications 1:30 pm: An image sensor with on-die diffractive optics in 0.18-micron bulk CMOS, C. J. Thomas, R. I. Hornsey, York Univ. (Canada) . . [6068-06] 1:50 pm: CMOS long linear array for space application, G. Lepage, Cypress Semiconductor Corp. (Belgium); D. G. Dantes, Alcatel Alenia Space (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-08] 2:10 pm: Experimental single-chip color HDTV image acquisition system with 8-Mpixel CMOS image sensor, H. Shimamoto, T. Yamashita, R. Funatsu, K. Mitani, Y. Nojiri, NHK Science & Technical Research Labs. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-09] 2:30 pm: High-sensitivity 2.5-μm pixel CMOS image sensor realized using Cu interconnect layers, K. Tatani, Y. Enomoto, A. Yamamoto, T. Goto, H. Abe, T. Hirayama, Sony Corp. (Japan) . . [6068-11] 2:50 pm: CMOS image sensor overlaid with organic photoelectric conversion layers and the proposal of stack type solid-state imaging devices, S. Takada, M. Inuiya, Y. Araki, Fuji Photo Film Co., Ltd. ( Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-12] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm 3:40 pm: An ultrawide dynamic-range CMOS image sensor with a linear response, J. H. Park, M. Mase, S. Kawahito, Shizuoka Univ. (Japan); M. Sasaki, Sendai National College of Technology (Japan); Y. Wakamori, Yamaha Corp. (Japan); Y. Ohta, Hamamatsu Corp. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-13] 4:00 pm: A pulse-frequency-modulation vision chip using a capacitive feedback reset with in-pixel 1-bit image processors, K. Kagawa, S. Yamamoto, T. Furumiya, T. Tokuda, M. Nunoshita, J. Ohta, Nara Institute of Science and Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . [6068-14]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
43
Conference 6068 · Conv. Ctr. Room A6
Thursday 19 January SESSION 3 Conv. Ctr. Room A6 . . . . . . . . . . . Thurs. 9:00 to 11:50 am Novel Devices and CCDs 9:00 am: Photon counting imaging: the DigitalAPD, S. Bellis, R. Wilcock, J. C. Jackson, SensL Technologies Ltd. (Ireland) . . . . . . . . . . . . . [6068-15] 9:20 am: Quantitative and qualitative performance comparison of a biomimetic vision sensor with commercial CCD camera sensors, R. S. Prabhakara, C. H. G. Wright, S. F. Barrett, W. M. Harman, Univ. of Wyoming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-16] 9:40 am: Quantum efficiency characterization of CCD's part 1: the quantum efficiency machine, D. E. Groom, C. J. Bebek, M. H. Fabricius, A. Karcher, J. Steckert, Lawrence Berkeley National Lab. . . . . . . [6068-35] 10:00 am: Quantum efficiency characterization of back-illuminated CCD's part 2: reflectivity measurements, M. H. Fabricius, Lawrence Berkeley National Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-36] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am 10:50 am: 28-M CCD imager with RGB compatible binning feature for professional applications, I. Peters, C. Draijer, F. Polderdijk, L. Meessen, B. G. M. Dillen, W. Klaassens, J. T. Bosiers, DALSA Professional Imaging (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-34] 11:10 am: Large area devices at Semiconductor Technology Associates, Inc., K. L. Boggs, R. A. Bredthauer, Semiconductor Technology Associates Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-17] 11:30 am: Development of the orthogonal-transfer array, B. E. Burke, MIT Lincoln Lab.; J. L. Tonry, Univ. of Hawai'i/West O'ahu; M. J. Cooper, MIT Lincoln Lab.; P. M. Onaka, Univ. of Hawai'i/West O'ahu; D. J. Young, A. H. Loomis, MIT Lincoln Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-18] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:50 am to 1:30 pm SESSION 4 Conv. Ctr. Room A6 . . . . . . . . . . . . Thurs. 1:30 to 4:40 pm Applications 1:30 pm: Toward 1-mm depth precision with a solid state full-field range imaging system, A. A. Dorrington, Univ. of Waikato (New Zealand); D. A. Carnegie, Victoria Univ. of Wellington (New Zealand); M. J. Cree, Univ. of Waikato (New Zealand) . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-22] 1:50 pm: An ultrafast phase modulator for 3D imaging, J. Y. Cheng, Northeastern Univ.; Q. Chen, Boston Applied Technologies Inc. . [6068-23] 2:10 pm: Classification of luminaire color using CCDs with application to airport lighting, K. R. McMenemy, J. H. Niblock, Queen's Univ. Belfast (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-24] 2:30 pm: Lateral chromatic aberration correction system for ultrahighdefinition color video camera, T. Yamashita, H. Shimamoto, R. Funatsu, K. Mitani, Y. Nojiri, NHK Science & Technical Research Labs. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-25] 2:50 pm: A new direct detection camera system for electron microscopy, N. Xuong, L. Jin, A. Milazzo, P. C. Leblanc, F. Duttweiler, J. C. Bouwer, S. Peltier, M. H. Ellisman, Univ. of California/San Diego; S. Li, S. Kleinfelder, Univ. of California/Irvine . . . . . . . . . . . . . . . . . . . . . . . [6068-26] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm 3:40 pm: A novel image processing system that autonomously monitors lighting patterns with application to airport lighting, J. H. Niblock, K. R. McMenemy, G. Irwin, Queen's Univ. Belfast (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-27] 4:00 pm: Wafer-scale UV embossing of aspheric lenses in phonecamera module, D. Shin, H. Jeong, S. Kim, S. C. Lee, Y. Jin, J. E. Noh, H. R. Oh, K. Lee, SAMSUNG Electro-Mechanics Co., Ltd. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6068-28] 4:20 pm: Colony optical image acquisition system, W. Wang, Chongqing Univ. of Posts and Telecommunications (China) . . . . . . . . . . . . . . [6068-29]
44
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6069 · Conv. Ctr. Room A6
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6069 Digital Photography II Conference Chairs: Nitin Sampat, Rochester Institute of Technology; Jeffrey M. DiCarlo, Hewlett-Packard Labs.; Russel A. Martin, Foveon USA Program Committee: Eiji Atsumi, Nokia Japan Co., Ltd. (Japan); Ted J. Cooper, Sony Electronics Inc.; Michael A. Kriss, Sharp Labs. of America; Jingqiang Li, Qualcomm, Inc.; Ricardo J. Motta, PIXIM, Inc.; Gloria G. Putnam, Eastman Kodak Co.; John R. Reinert Nash, Lifetouch, Inc.; Brian G. Rodricks, Micron Technology, Inc.; Sabine E. Sьsstrunk, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland)
Monday 16 January SESSION 1 Conv. Ctr. Room A6 . . . . . . . . . . . . Mon. 8:30 to 10:20 am Sensor Design Chair: Russel A. Martin, Foveon, Inc. 8:30 am: A brief history of `pixel' (Invited Paper), R. F. Lyon, Foveon, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-01] 9:00 am: 31 Mp and 39 Mp full-frame CCD image sensors with improved charge capacity and angle response, E. J. Meisenzahl, E. K. Banghart, D. N. Nichols, J. P. Shepherd, E. G. Stevens, K. Y. Wong, Eastman Kodak Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-02] 9:20 am: Improving low-light CMOS performance with four-transistor four-shared pixel architecture and charge-domain binning, F. Chu, R. M. Guidash, J. Compton, S. Coppola, W. Hintz, Eastman Kodak Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-03] 9:40 am: Optical interaction of space and wavelength in highresolution digital imagers, B. G. Rodricks, K. Venkataraman, Micron Technology, Inc.; P. B. Catrysse, B. A. Wandell, Stanford Univ. . . [6069-04] 10:00 am: Image recovery for a direct color imaging approach using a color filter array, T. Saito, T. Komatsu, Kanagawa Univ. (Japan) [6069-05] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am SESSION 2 Conv. Ctr. Room A6 . . . . . . . . Mon. 10:50 am to 12:10 pm Demosaicing Chair: Michael A. Kriss, Consultant 10:50 am: Demosaicing: heterogeneity projection hard-decision adaptive interpolation using spectral-spatial correlation, C. Tsai, K. Song, National Chiao Tung Univ. (Taiwan) . . . . . . . . . . . . . . . . . . [6069-06] 11:10 am: Iterative asymmetric average interpolation for color demosaicing of single-sensor digital camera data, Y. Takahashi, H. Kikuchi, S. Muramatsu, Niigata Univ. (Japan); N. Mizutani, Kodak Japan Ltd. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-07] 11:30 am: Spatially adaptive superresolution sharpeningdemosaicking for a single solid state color image sensor, T. Saito, T. Komatsu, Kanagawa Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . [6069-08] 11:50 am: Generic MSFA mosaicing and demosaicing for multispectral cameras, L. Miao, H. Qi, The Univ. of Tennessee; R. Ramanath, North Carolina State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-09] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:40 pm
SESSION 3 Conv. Ctr. Room A6 . . . . . . . . . . . . . Mon. 1:40 to 3:00 pm Auto Exposure, Focus, and White Balance Chair: Jeffrey M. DiCarlo, Hewlett-Packard Labs. 1:40 pm: Dynamic focus window selection using a statistical color model, Y. Tian, Univ. of California/Berkeley . . . . . . . . . . . . . . . . . [6069-10] 2:00 pm: Combinational AE-AF system with fuzzy climbing search servo, C. Chen, C. Tseng, C. Hung, I. Yin, S. Wang, National Chiao Tung Univ. (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-11] 2:20 pm: Multidomain pixel analysis for illuminant estimation, F. Gasparini, R. Schettini, Univ. degli Studi di Milano-Bicocca (Italy); F. Naccari, A. Bruna, STMicroelectronics (Italy) . . . . . . . . . . . . . . . . [6069-12] 2:40 pm: Computational inexpensive two-step auto white balance method, S. R. Goma, M. Aleksic, ATI Technology (Canada) . . . . [6069-13] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm SESSION 4 Conv. Ctr. Room A6 . . . . . . . . . . . . . Mon. 3:30 to 4:50 pm Image Enhancement Chair: Brian G. Rodricks, Micron Technology, Inc. 3:30 pm: An effective image enhancement filtering for noisy image sequences, H. Lee, D. Park, S. Lee, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) . . . . . . . . . . . . . . . . . . . . . . [6069-14] 3:50 pm: Novel bilateral filter approach: image noise reduction with sharpening, M. Aleksic, M. Smirnov, S. R. Goma, ATI Technology (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-15] 4:10 pm: Digital photograph stitching with optimized matching of gradient and curvature, S. T. Suen, E. Y. Lam, K. K. Wong, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-16] 4:30 pm: Compensation of nonuniform flash illumination in group portrait photography, J. H. Kim, Pukyong National Univ. (South Korea); B. A. Barsky, Univ. of California/Berkeley . . . . . . . . . . . . . . . . . . . . . [6069-17] SESSION 5 Conv. Ctr. Room A6 . . . . . . . . . . . . . Mon. 4:50 to 5:30 pm Imaging Systems Chair: John R. Reinert Nash, Lifetouch, Inc. 4:50 pm: A robotic system for digital photography, L. W. MacDonald, London College of Communication . . . . . . . . . . . . . . . . . . . . . . . [6069-18] 5:10 pm: Source camera identification using footprints from JPEG compression and lens aberration, K. S. Choi, E. Y. Lam, K. K. Wong, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . [6069-19]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
45
Conference 6069 · Conv. Ctr. Room A6 Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 6 Conv. Ctr. Room A6 . . . . . . . . . . . . Tues. 9:30 to 10:20 am Camera Evaluation I Chair: Nitin Sampat, Rochester Institute of Technology 9:30 am: Evaluating digital cameras (Invited Paper), D. Wueller, Image Engineering (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-20] 10:00 am: The Imatest program: comparing cameras with different amounts of sharpening, N. L. Koren, Imatest LLC . . . . . . . . . . . [6069-21] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am SESSION 7 Conv. Ctr. Room A6 . . . . . . . . Tues. 10:50 am to 12:10 pm Camera Evaluation II Chair: Jingqiang Li, Rochester Institute of Technology 10:50 am: Resolution for color photography, P. M. Hubel, Foveon, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-22] 11:10 am: Resolution and light sensitivity tradeoff with pixel size, J. E. Farrell, Stanford Univ.; F. Xiao, Agilent Technologies; S. Kavusi, Stanford Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6069-23] 11:30 am: Characterization of noise in digital photographs for image processing, S. Lim, Hewlett-Packard Co. . . . . . . . . . . . . . . . . . . [6069-24] 11:50 am: Proposal for a standard procedure to test mobile phone cameras, D. Wueller, Image Engineering (Germany) . . . . . . . . . . [6069-25] Demonstration Session A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging.
46
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6070 · Conv. Ctr. Room C3
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6070 Machine Vision Applications in Industrial Inspection XIV Conference Chairs: Fabrice Meriaudeau, Univ. de Bourgogne (France); Kurt S. Niel, Fachhochschule Wels (Austria) Program Committee: Pierrick T. Bourgeat, BioMedIA Lab. (Australia); Luciano da Fontoura Costa, Univ. de Sгo Paulo (Brazil); Marc M. Ellenrieder, Daimler Chrysler AG (Germany); Steven P. Floeder, 3M Co.; David Fofi, Univ. de Bourgogne (France); Ralph M. Ford, The Pennsylvania State Univ.; Edmund Y. Lam, The Univ. of Hong Kong (Hong Kong China); Katia Lebart, Heriot-Watt Univ. (United Kingdom); Dinesh Nair, National Instruments; Paul L. O'Leary, Montan Univ. Leoben (Austria); Jeffery R. Price, Oak Ridge National Lab.; A. Ravishankar Rao, IBM Thomas J. Watson Research Ctr.; Joaquim Salvi, Univ. de Girona (Spain); Hamed Sari-Sarraf, Texas Tech Univ.; Christoph Stiller, Univ. Karlsruhe (Germany); Kenneth W. Tobin, Jr., Oak Ridge National Lab.; Yvon Voisin, Univ. de Bourgogne (France)
Monday 16 January SESSION 1 Conv. Ctr. Room C3 . . . . . . . . . . . . Mon. 8:40 to 10:00 am Industrial Applications I Chair: Fabrice Meriaudeau, Univ. de Bourgogne (France) 8:40 am: Optical servoing for industrial surface machining, N. Koller, Hotvision Research GmbH (Austria); R. Ofner, P. L. O'Leary, Montan Univ. Leoben (Austria); E. Fauster, Hotvision Research GmbH (Austria) [6070-01] 9:00 am: Statistical learning with imbalanced training set in a machine vision application: improve the false alarm rate and sensitivity simultaneously, J. Q. Li, Agilent Technologies, Inc. . . . . . . . . . . [6070-02] 9:20 am: Boundary detection of projected fringes for surface of inhomogeneous reflectance, J. Cheng, R. C. Chung, The Chinese Univ. of Hong Kong (Hong Kong China); E. Y. Lam, The Univ. of Hong Kong (Hong Kong China); K. S. M. Fung, F. Wang, W. H. Leung, ASM Assembly Automation Ltd. (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . [6070-03] 9:40 am: Height inspection of wafer bumps without explicit 3D reconstruction, M. Dong, R. C. Chung, Y. Zhao, The Chinese Univ. of Hong Kong (Hong Kong China); E. Y. Lam, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-04] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am SESSION 2 Conv. Ctr. Room C3 . . . . . . . . Mon. 10:30 am to 12:10 pm Multispectral Imaging Chair: Kurt S. Niel, Fachhochschule Wels (Austria) 10:30 am: Fast recognition method for metallic topographies by the "Three-Color Selective Stereo Gradient Method" (Three-Color SSGM), M. Hossfeld, Technische Univ. Hamburg-Harburg (Germany); M. Adameck, Hella KGaA Hueck & Co. (Germany); M. Eich, Technische Univ. HamburgHarburg (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-05] 10:50 am: Robustness of texture parameters for color texture analysis, A. Marin, A. Roman, Univ. de Bourgogne (France); D. R. Connah, J. Y. Hardeberg, Gjшvik Univ. College (Norway); P. Gouton, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-06] 11:10 am: Real-time multispectral imaging application for poultry safety inspection, B. Park, K. C. Lawrence, W. R. Windham, M. P. Snead, U.S. Dept. of Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-07] 11:30 am: Flushing analysis by machine vision and fuzzy logic at molten steel for the automation process, C. Pfob, K. S. Niel, Fachhochschule Wels (Austria); R. Rцssler, voestalpine Mechatronics GmbH (Austria) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-08] 11:50 am: Color influence in accuracy of 3D scanner based on structured light, S. Voisin, The Univ. of Tennessee and Univ. de Bourgogne (France); D. L. Page, The Univ. of Tennessee; S. Foufou, F. Truchetet, Univ. de Bourgogne (France); M. A. Abidi, The Univ. of Tennessee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-09] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:40 pm
SESSION 3 Conv. Ctr. Room C3 . . . . . . . . . . . . . Mon. 1:40 to 2:40 pm 3D Applications I Chair: Jeffery R. Price, Oak Ridge National Lab. 1:40 pm: Novel view synthesis for projective texture mapping on real 3D objects, T. Molinier, D. Fofi, P. Gorria, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-10] 2:00 pm: 3D translucent object reconstruction from artificial vision, F. Truchetet, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . [6070-11] 2:20 pm: Real-time 3D wood panel surface measurement using laser triangulation and low-cost hardware, H. Ramoser, L. Cambrini, H. Rцtzer, Advanced Computer Vision (Austria) . . . . . . . . . . . . . . . . [6070-12] SESSION 4 Conv. Ctr. Room C3 . . . . . . . . . . . . . Mon. 2:40 to 4:50 pm Industrial Applications II Chair: Kurt S. Niel, Fachhochschule Wels (Austria) 2:40 pm: Aerial platform attitude measurement by artificial vision, F. Truchetet, O. Aubreton, P. Gorria, O. Laligant, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-14] 3:00 pm: A new algorithm for real-time multistage image thresholding, S. H. Lin, R. Giesen, D. Nair, National Instruments . . . . . . . . . . . . [6070-15] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:20 to 3:50 pm 3:50 pm: Simultaneous photometric correction and defect detection in semiconductor manufacturing, Y. Shen, E. Y. Lam, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-16] 4:10 pm: Automatic mura detection system for liquid crystal display panels, L. Fang, H. Chen, I. Yin, S. Wang, C. Wen, C. Kuo, National Chiao Tung Univ. (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-17] 4:30 pm: New developments in image-based characterization of coated particle nuclear fuel, J. R. Price, D. Aykac, J. D. Hunn, A. K. Kercher, R. N. Morris, Oak Ridge National Lab. . . . . . . . . . . . . . . [6070-18]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
47
Conference 6070 · Conv. Ctr. Room C3
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 5 Conv. Ctr. Room C3 . . . . . . . . . . . . Tues. 9:30 to 11:40 am Multiresolution and Mathematical Fitting Chair: Fabrice Meriaudeau, Univ. de Bourgogne (France) 9:30 am: Real-time detection of elliptic shapes for automated object recognition and object tracking, C. Teutsch, D. Berndt, E. Trostmann, M. Weber, Fraunhofer-Institut fьr Fabrikbetrieb und -automatisierung (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-19] 9:50 am: Discrete circles measurement for industrial inspection, F. Mairesse, T. M. Sliwa, S. Binczak, Y. Voisin, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-20] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am 10:40 am: Twin and scratch detection and removal in micrograph images of Inconel 718, G. Jakob, A. Rinnhofer, Joanneum Research (Austria); H. Bischof, Technische Univ. Graz (Austria); W. Benesova, Joanneum Research (Austria) . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-21] 11:00 am: Tracking fluorescent spots in wide-field microscopy images, L. A. Muresan, B. Heise, E. P. Klement, Johannes Kepler Univ. Linz (Austria) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-22] 11:20 am: Development of method based on Hough transform or Gabor filtering to discriminate crop and weeds in agronomic image, J. Bossu, C. Gee, J. Guillemin, Etablissement National d'Enseignement Superieur Agronomique de Dijon (France); F. Truchetet, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-23] SESSION 6 Conv. Ctr. Room C3 . . . . . . . . Tues. 11:40 am to 12:20 pm 3D Applications II Chairs: Fabrice Meriaudeau, Univ. de Bourgogne (France); Kurt S. Niel, Fachhochschule Wels (Austria) 11:40 am: A refined range image registration technique applied to multistripe laser 3D scanner, C. Matabosch, J. Salvi, Univ. de Girona (Spain); D. Fofi, F. Meriaudeau, Univ. de Bourgogne (France) . . . [6070-24] 12:00 pm: Surface orientation recovery of specular microsurface via binary pattern projection, Z. Song, R. C. Chung, J. Cheng, The Chinese Univ. of Hong Kong (Hong Kong China); E. Y. Lam, The Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-25]
Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Constructing a simple parametric model of shoulder from medical images, H. Atmani, F. Mйrienne, Йcole Nationale Supйrieure d'Arts et Mйtiers (France); D. Fofi, P. Trouilloud, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-26] A study of automatic monitoring and measuring vehicles by using image analysis, W. Wang, Chongqing Univ. of Posts and Telecommunications (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-27] An active contour algorithm for detecting the circular features in a PCB x-ray image, Y. Chen, C. Wu, W. Hu, Yuan Ze Univ. (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-28] Human vision based detection of nonuniform brightness on LCD panels, J. H. Kim, Pukyong National Univ.; B. A. Barsky, Univ. of California/Berkeley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-30] Optimized texture classification by using hierarchical complex networks, T. Chalumeau, Univ. de Bourgogne (France); L. F. da Fontoura Costa, Univ. de Sгo Paulo (Brazil); F. Meriaudeau, O. Laligant, Univ. de Bourgogne (France) . . . . . . . . . . . . . . . . . . . . . . . . . . [6070-31]
48
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6071 · Conv. Ctr. Room C1
Wednesday-Thursday 18-19 January 2006 · Proceedings of SPIE Vol. 6071 Multimedia Computing and Networking 2006
Conference Chairs: Surendar Chandra, Univ. of Notre Dame; Carsten Griwodz, Univ. of Oslo (Norway) Program Committee: Tarek F. Abdelzaher, Univ. of Virginia; Sarita V. Adve, Univ. of Illinois at Urbana-Champaign; Scott A. Brandt, Univ. of California/Santa Cruz; David H. Du, Univ. of Minnesota; Wu-chi Feng, Portland State Univ.; Pascal Frossard, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland); Pеl Halvorsen, Simula Research Lab. (Norway); Baochun Li, Univ. of Toronto (Canada); Ian Marsh, Swedish computer science Institute (Sweden); Andreas U. Mauthe, Lancaster Univ. (United Kingdom); Ketan D. Mayer-Patel, The Univ. of North Carolina at Chapel Hill; Klara Nahrstedt, Univ. of Illinois at Urbana-Champaign; Wei-Tsang Ooi, National Univ. of Singapore (Singapore); Ragunathan Rajkumar, Carnegie Mellon Univ.; Karsten Schwan, Georgia Institute of Technology; Tajana Simunic-Rosing, Univ. of California/ San Diego; Ralf Steinmetz, Technische Univ. Darmstadt (Germany); Nalini Venkatasubramanian, Univ. of California/ Irvine; Lars Wolf, Technische Univ. Braunschweig (Germany); Dongyan Xu, Purdue Univ.; Wanghong Yuan, DoCoMo Communications Labs. USA, Inc.; Roger Zimmermann, Univ. of Southern California; Michael Zink, Univ. of Massachusetts/Amherst
Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 1 Conv. Ctr. Room C1 . . . . . . . . . Wed. 9:30 am to 12:30 pm Application-dependent Transfer 9:30 am: The effects of frame rate and resolution on users playing first person shooter games, M. Claypool, Worcester Polytechnic Institute; K. Claypool, F. Damaa, Univ. of Massachusetts/Lowell . . . . . . . . . . [6071-01] 10:00 am: Real-time 3D video compression for tele-immersive environments, Z. Yang, Y. Cui, Z. Anwar, R. Bocchino, N. Kiyanclar, K. Nahrstedt, R. H. Campbell, Univ. of Illinois at Urbana-Champaign; W. J. Yurcik, National Ctr. for Supercomputing Applications . . . . . . . . [6071-02] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am 11:00 am: An integrated visual approach for music indexing and dynamic playlist composition, M. Crampes, S. Ranwez, Ecole des Mines d'Alиs (France); F. Velickovski, C. Mooney, Univ. of New South Wales (Australia); N. Mille, Netia, Inc. (France) . . . . . . . . . . . . . . . . . . . . [6071-03] 11:30 am: Efficient rate-distortion optimized media streaming for treereducible packet dependencies, M. Rцder, Univ. of Konstanz (Germany); J. Cardinal, Univ. Libre de Bruxelles (Belgium); R. Hamzaoui, Univ. of Konstanz (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-04] 12:00 pm: Popular song and lyrics synchronization and its application to music information retrieval, K. Chen, S. Gao, Y. Zhu, Q. Sun, Institute for Infocomm Research (Singapore) . . . . . . . . . . . . . . . . . . . . . . . [6071-05] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:30 to 2:00 pm SESSION 2 Conv. Ctr. Room C1 . . . . . . . . . . . . . Wed. 2:00 to 4:30 pm Streaming 2:00 pm: MMS: a multihome-aware media streaming system, A. Habib, Siemens Technology to Business Ctr.; J. Chuang, Univ. of California/ Berkeley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-06] 2:30 pm: Streamline: a scheduling heuristic for streaming applications on the grid, B. Agarwalla, N. Ahmed, D. Hilley, U. Ramachandran, Georgia Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-07] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm
3:30 pm: A novel unbalanced multiple description coder for robust video transmission over ad hoc wireless networks, F. Huang, L. Sun, Y. Zhong, Tsinghua Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-08] 4:00 pm: A transform for network calculus and its application to multimedia networking, K. Pandit, Technische Univ. Darmstadt (Germany); J. Schmitt, C. Kirchner, Univ. Kaiserslautern (Germany); R. Steinmetz, Technische Univ. Darmstadt (Germany) . . . . . . . . . . . [6071-09] Panel Discussion . . . . . . . . . . . . . . . . . . . 4:30 to 5:30 pm Multimedia Sensors: Technological and Societal Challenges Chair: Roger Zimmermann, Univ. of Southern California Thursday 19 January
SESSION 3 Conv. Ctr. Room C1 . . . . . . . . . . . Thurs. 8:30 to 10:30 am Distribution 8:30 am: A method to deliver multi-object content in a ubiquitous environment, T. Mori, M. Katsumoto, National Institute of Information and Communications Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . [6071-10] 9:00 am: Correlation-aware multimedia content distribution in overlay networks, Y. Zhu, B. Li, Univ. of Toronto (Canada) . . . . . . . . . . . [6071-11] 9:30 am: QBIX-G: a transcoding multimedia proxy, P. Schojer, L. Boeszoermenyi, H. Hellwagner, Univ. Klagenfurt (Austria) . . . . . . [6071-12] 10:00 am: Preventing DoS attacks in peer-to-peer media streaming systems, W. G. Conner II, K. Nahrstedt, I. Gupta, Univ. of Illinois at Urbana-Champaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-13] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am
SESSION 4
Conv. Ctr. Room C1 . . . . . . . Thurs. 11:00 am to 12:00 pm
Keynote
Keynote Presentation
11:00 am: Playstation and multimedia (Invited Paper, Presentation Only), K. Hofrichter, Sony Computer Entertainment America . . . [6071-14]
Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
49
Conference 6071 · Conv. Ctr. Room C1 SESSION 5 Conv. Ctr. Room C1 . . . . . . . . . . . . Thurs. 1:30 to 3:00 pm Short Papers: Multimedia Systems 1:30 pm: Investigating a stream synchronization middleware for the NEES MAST system, J. C. Beyer, S. K. Chirravuri, D. H. Du, Univ. of Minnesota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-15] 1:45 pm: A performance model of effective memory management in HYDRA: a large-scale data stream recording system, K. Fu, R. Zimmermann, Univ. of Southern California . . . . . . . . . . . . . . . . . . [6071-16] 2:00 pm: Sender-driven bandwidth differentiation for transmitting multimedia flows over TCP, J. K. H. Lau, J. Y. B. Lee, The Chinese Univ. of Hong Kong (Hong Kong China) . . . . . . . . . . . . . . . . . . . . . . . . [6071-17] 2:15 pm: FlexSplit: a workload-aware adaptive load balancing strategy for media clusters, Q. Zhang, College of William & Mary; L. Cherkasova, Hewlett-Packard Labs.; E. Smirni, College of William & Mary . . . [6071-18] 2:30 pm: Cascades: scalable, flexible, and composable middleware for multimodal sensor networking applications, J. Huang, W. Feng, N. Bulusu, W. Feng, Portland State Univ. . . . . . . . . . . . . . . . . . . . . . [6071-19] 2:45 pm: Compression by indexing: an improvement over MPEG-4 body animation parameter compression, S. Chattopadhyay, S. M. Bhandarkar, K. Li, The Univ. of Georgia . . . . . . . . . . . . . . . . . . . . [6071-20] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm SESSION 6 Conv. Ctr. Room C1 . . . . . . . . . . . . Thurs. 3:30 to 5:30 pm Peer-to-Peer 3:30 pm: DagStream: locality aware and failure resilient peer-to-peer streaming, J. Liang, K. Nahrstedt, Univ. of Illinois at UrbanaChampaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-21] 4:00 pm: Characterizing files in the modern Gnutella network: a measurement study, S. Zhao, D. Stutzbach, R. Rejaie, Univ. of Oregon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-22] 4:30 pm: Sampling cluster endurance for peer-to-peer based content distribution networks, V. Darlagiannis, Technische Univ. Darmstadt (Germany); A. U. Mauthe, Lancaster Univ. (United Kingdom); R. Steinmetz, Technische Univ. Darmstadt (Germany) . . . . . . . . . . . . . . . . . . . . [6071-23] 5:00 pm: How efficient is BitTorrent?, G. Wu, T. Chiueh, Stony Brook Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6071-24]
50
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6072 · Conv. Ctr. Room A5
Monday-Thursday 16-19 January 2006 · Proceedings of SPIE Vol. 6072 Security, Steganography, and Watermarking of Multimedia Contents VIII
Conference Chairs: Edward J. Delp III, Purdue Univ.; Ping Wah Wong, Nethra Imaging Program Committee: Adnan M. Alattar, Digimarc Corp.; Mauro Barni, Univ. degli Studi di Siena (Italy); Jeffrey A. Bloom, Thomson Multimedia Corporate Research; Gordon W. Braudaway, IBM Corp.; Ingemar J. Cox, Univ. College London (United Kingdom); Jana Dittmann, Otto-von-Guericke-Univ. Magdeburg (Germany); Ahmet M. Eskicioglu, The City Univ. of New York; Jessica Fridrich, Binghamton Univ.; Teddy Furon, IRISA (France); Ton Kalker, Hewlett-Packard Co.; Martin Kutter, AlpVision SA (Switzerland); Reginald L. Lagendijk, Technische Univ. Delft (Netherlands); Benoоt B. Macq, Univ. Catholique de Louvain (Belgium); Bangalore S. Manjunath, Univ. of California/Santa Barbara; Nasir D. Memon, Polytechnic Univ.; Pierre Moulin, Univ. of Illinois at Urbana-Champaign; Fernando PйrezGonzбlez, Univ. de Vigo (Spain); Gaurav Sharma, Univ. of Rochester; Claus Vielhauer, Otto-von-Guericke-Univ. Magdeburg (Germany); Sviatoslav V. Voloshynovskiy, Univ. de Genиve (Switzerland); Min Wu, Univ. of Maryland/College Park
Monday 16 January SESSION 1 Conv. Ctr. Room A5 . . . . . . . . . . . . Mon. 8:30 to 11:20 am Steganography and Steganalysis I Chair: Jessica Fridrich, Binghamton Univ. 8:30 am: New blind steganalysis and its implications, M. Goljan, J. Fridrich, T. S. Holotyak, Binghamton Univ. . . . . . . . . . . . . . . . . . . [6072-01] 8:50 am: Statistical modeling and steganalysis of DFT-based image steganography, Y. Wang, P. Moulin, Univ. of Illinois at UrbanaChampaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-02] 9:10 am: Fourth-order structural steganalysis and analysis of cover assumptions, A. D. Ker, Univ. of Oxford (United Kingdom) . . . . . [6072-03] 9:30 am: Application of conditional entropy measures to steganalysis, J. A. Marsh, T. Knapik, E. Lo, SI International; C. D. Heitzenrater, Air Force Research Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-04] 9:50 am: Improving steganalysis by fusion techniques: a case study with image-based steganography, M. Kharrazi, T. H. Sencar, N. D. Memon, Polytechnic Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-05] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am 10:40 am: A two-factor error model for quantitative steganalysis, R. Bцhme, Technische Univ. Dresden (Germany); A. D. Ker, Univ. of Oxford (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-06] 11:00 am: Compression-based steganalysis of LSB embedded images, C. G. Boncelet, Jr., Univ. of Delaware; L. M. Marvel, A. J. Raglin, Army Research Lab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-07] SESSION 2 Conv. Ctr. Room A5 . . . . . . . . Mon. 11:20 am to 12:20 pm Special Session: Natural Language Watermarking Chairs: Mercan Topkara, Purdue Univ.; Cuneyt M. Taskiran, Motorola, Inc. 11:20 am: Natural language processing with linguistic information for digital fingerprinting and watermarking, O. Uzuner, Massachusetts Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-08] 11:40 am: Attacks on linguistic steganography systems using text analysis, C. M. Taskiran, Motorola, Inc.; M. Topkara, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-09] 12:00 pm: Natural language watermarking: research challenges and applications, M. Topkara, Purdue Univ.; G. Riccardi, D. Hakkani-Tьr, AT&T Labs. Research; M. J. Atallah, Purdue Univ. . . . . . . . . . . . . . . . . . [6072-10] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:20 to 1:50 pm
SESSION 3 Conv. Ctr. Room A5 . . . . . . . . . . . . . Mon. 1:50 to 3:30 pm Attacks Chair: Fernando Pйrez-Gonzбlez, Univ. de Vigo (Spain) 1:50 pm: Scale estimation in two-band filter attacks on QIM watermarks, J. Wang, Technische Univ. Delft (Netherlands) and Nanjing Univ. (China); I. D. Shterev, R. L. Lagendijk, Technische Univ. Delft (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-11] 2:10 pm: High-rate quantization data hiding robust to arbitrary linear filtering attacks, F. Pйrez-Gonzбlez, C. Mosquera, M. Alvarez-Diaz, Univ. de Vigo (Spain); R. L. Lagendijk, Technische Univ. Delft (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-12] 2:30 pm: Countermeasure for collusion attacks against digital watermarking, M. Steinebach, S. Zmudzinski, Fraunhofer-Institut fьr Integrierte Publikations- und Informationssysteme (Germany) . . . [6072-13] 2:50 pm: The blind Newton sensitivity attack, P. Comesaсa Alfaro, L. L. Pйrez-Freire, F. Pйrez-Gonzбlez, Univ. de Vigo (Spain) . . . . . . . . . [6072-14] 3:10 pm: Achieving non-ambiguity of quantization-based watermarking, X. Kang, Sun Yat-Sen Univ. (China) and New Jersey Institute of Technology; Y. Shi, New Jersey Institute of Technology; J. Huang, Sun Yat-Sen Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . [6072-15] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:30 to 4:00 pm SESSION 4 Conv. Ctr. Room A5 . . . . . . . . . . . . . Mon. 4:00 to 5:40 pm Special Session: Biometrics Chair: Claus Vielhauer, Otto-von-Guericke-Univ. Magdeburg (Germany) 4:00 pm: Reference point detection for improved fingerprint matching, T. Ignatenko, Technische Univ. Eindhoven (Netherlands); T. Kalker, HewlettPackard Co. and Technische Univ. Eindhoven (Netherlands); M. van der Veen, Philips Research Labs. (Netherlands); A. M. Bazen, Univ. Twente (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-65] 4:20 pm: Analyzing handwriting biometrics in metadata context, F. Wolf, T. Scheidat, C. Vielhauer, Otto-von-Guericke-Univ. Magdeburg (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-17] 4:40 pm: 3D face recognition by projection-based features, B. Sankur, H. Dutagaci, Bogaziзi Univ. (Turkey); Y. Yemez, Koз Univ. (Turkey) [6072-18] 5:00 pm: Face biometrics with renewable templates, M. van der veen, T. Kevenaar, T. H. Akkermans, G. Schrijen, Philips Research Labs. (Netherlands); F. Zuo, Technische Univ. Eindhoven (Netherlands) [6072-19] 5:20 pm: Safety of templates in biometric person authentication using error-correcting code, T. Ohki, S. Akatsuka, N. Komatsu, Waseda Univ. (Japan); M. Kasahara, Osaka Gakuin Univ. (Japan) . . . . . . . . . . . [6072-20]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
51
Conference 6072 · Conv. Ctr. Room A5
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 5 Conv. Ctr. Room A5 . . . . . . . . . . . . Tues. 9:30 to 10:10 am Audio Chair: Scott A. Craver, Binghamton Univ. 9:30 am: On the comparison of audio fingerprints for extracting quality parameters of compressed audio, P. J. Doets, M. Menor Gisbert, R. L. Lagendijk, Technische Univ. Delft (Netherlands) . . . . . . . . . . . . . . [6072-21] 9:50 am: Fingerprinting with Wow, S. A. Craver, Binghamton Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-22] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am SESSION 6 Conv. Ctr. Room A5 . . . . . . . . Tues. 10:40 am to 12:00 pm Steganography and Steganalysis II Chair: Benoоt B. Macq, Univ. Catholique de Louvain (Belgium) 10:40 am: Limited distortion in LSB steganography, Y. Kim, Z. Duric, D. Richards, George Mason Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-23] 11:00 am: Multiclass blind steganalysis for JPEG images, J. Fridrich, T. Pevny, Binghamton Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-24] 11:20 am: MPsteg: hiding a message in the matching pursuit domain, G. Cancelli, M. Barni, G. Menegaz, Univ. degli Studi di Siena (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-25] 11:40 am: Stego sensitivity measure and multibit plane based steganography using different color models, S. S. Agaian, J. P. Perez, B. M. Rodriguez II, The Univ. of Texas at San Antonio . . . . . . . . . [6072-26] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm SESSION 7 Conv. Ctr. Room A5 . . . . . . . . . . . . . Tues. 1:30 to 4:20 pm Embedding I Chair: Ping Wah Wong, IDzap LLC 1:30 pm: Zero knowledge ST-DM watermarking, A. Piva, D. Corazzi, A. De Rosa, Univ. degli Studi di Firenze (Italy); M. Barni, Univ. degli Studi di Siena (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-27] 1:50 pm: Compression and rotation resistant watermark using a circular chirp structure, C. E. Fleming, B. G. Mobasseri, Villanova Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-28] 2:10 pm: Rotation/scale insensitive spread spectrum image watermarking game, M. Ossonce, G. Le Guelvouit, C. Delpha, P. Duhamel, Lab. des signaux et systиmes (France) . . . . . . . . . . . . [6072-29] 2:30 pm: New results on robustness of secure steganography, M. T. Hogan, F. Balado, N. J. Hurley, G. C. M. Silvestre, National Univ. of Ireland/ Dublin (Ireland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-30] 2:50 pm: Sphere-hardening dither modulation, F. Balado, N. J. Hurley, G. C. M. Silvestre, National Univ. of Ireland/Dublin (Ireland) . . . . . . . [6072-31] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm
3:40 pm: Secret dither estimation in lattice-quantization data hiding: a set membership approach, L. L. Pйrez-Freire, F. Pйrez-Gonzбlez, P. Comesaсa Alfaro, Univ. de Vigo (Spain) . . . . . . . . . . . . . . . . . . . . [6072-32] 4:00 pm: Performance analysis of nonuniform quantization-based data hiding, J. E. Vila-Forcйn, S. V. Voloshynovskiy, O. J. Koval, T. Pun, Univ. de Genиve (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-33] SESSION 8 Conv. Ctr. Room A5 . . . . . . . . . . . . . Tues. 4:20 to 5:40 pm Special Session: Forensics Chair: Nasir D. Memon, Polytechnic Univ. 4:20 pm: Detecting digital image forgeries using sensor pattern noise, J. Lukas, J. Fridrich, M. Goljan, Binghamton Univ. . . . . . . . . . . . . [6072-34] 4:40 pm: Fingerprinting digital elevation maps, H. Gou, M. Wu, Univ. of Maryland/College Park . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-35] 5:00 pm: Information embedding and extraction for electrophotographic printing processes, A. K. Mikkilineni, P. Chiang, G. T. Chiu, J. P. Allebach, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . [6072-36] 5:20 pm: An online system for classifying computer graphics images from natural photographs, T. Ng, S. Chang, Columbia Univ. . . . [6072-37] Demonstration Session A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 9 Conv. Ctr. Room A5 . . . . . . . . . . . . Wed. 9:30 to 11:20 am Theoretical Methods I Chair: Pierre Moulin, Univ. of Illinois at Urbana-Champaign 9:30 am: Text data-hiding for digital and printed documents: theoretical and practical considerations, R. Villan, Sr., S. V. Voloshynovskiy, O. J. Koval, J. E. Vila-Forcйn, E. Topak, F. Deguillaume, Y. B. Rytsar, T. Pun, Univ. de Genиve (Switzerland) . . . . . . . . . . . . . [6072-38] 9:50 am: E-capacity and security analysis of data-hiding channels with geometrical attacks, E. Topak, S. V. Voloshynovskiy, O. J. Koval, Univ. de Genиve (Switzerland); M. E. Haroutunian, National Academy of Sciences of Armenia (Armenia); J. E. Vila-Forcйn, T. Pun, Univ. de Genиve (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-39] 10:10 am: Image data hiding based on capacity-approaching dirtypaper coding, Y. Yang, Y. Sun, V. Stankovic, Z. Xiong, Texas A&M Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-40] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am 11:00 am: Wet paper codes with improved embedding efficiency, J. Fridrich, M. Goljan, D. Soukal, Binghamton Univ. . . . . . . . . . . . . [6072-41]
52
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6072 · Conv. Ctr. Room A5
SESSION 10 Conv. Ctr. Room A5 . . . . . . . . Wed. 11:20 am to 12:20 pm Video I Chair: Adnan M. Alattar, Digimarc Corp. 11:20 am: New modulation-based watermarking technique for video, A. N. Lemma, M. van der Veen, M. U. Celik, Philips Research Labs. (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-42] 11:40 am: Selective encryption for H.264/AVC video coding, T. Shi, B. King, P. Salama, Indiana Univ./Purdue Univ. at Indianapolis . . . . [6072-43] 12:00 pm: Using entropy for image and video authentication watermarks, S. Thiemert, M. Steinebach, Fraunhofer-Institut fьr Integrierte Publikations- und Informationssysteme (Germany) . . . . . . . . . . . [6072-44] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:20 to 1:50 pm SESSION 11 Conv. Ctr. Room A5 . . . . . . . . . . . . . Wed. 1:50 to 3:10 pm Video II Chair: Ahmet M. Eskicioglu, The City Univ. of New York 1:50 pm: Temporal synchronization of marked MPEG video frames based on image hash system, E. Hauer, M. Steinebach, FraunhoferInstitut fьr Integrierte Publikations- und Informationssysteme (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-45] 2:10 pm: Towards robust compressed-domain video watermarking for H.264, M. Noorkami, R. M. Mersereau, Georgia Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-46] 2:30 pm: Selective encryption of low-complexity source coding for mobile terminals, H. Um, E. J. Delp III, Purdue Univ. . . . . . . . . . [6072-47] 2:50 pm: VLC pair tree: a paradigm for MPEG-2 watermarking, M. P. Marcinak, B. G. Mobasseri, Villanova Univ. . . . . . . . . . . . . . . . . . [6072-48] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm SESSION 12 Conv. Ctr. Room A5 . . . . . . . . . . . . . Wed. 3:40 to 5:40 pm Theoretical Methods II Chair: Fernando Pйrez-Gonzбlez, Univ. de Vigo (Spain) 3:40 pm: Information-theoretic analysis of electronic and printed document authentication, S. V. Voloshynovskiy, O. J. Koval, R. Villan, Sr., E. Topak, J. E. Vila-Forcйn, F. Deguillaume, Y. B. Rytsar, T. Pun, Univ. de Genиve (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-49] 4:00 pm: Joint data hiding and source coding with partially available side information, C. Dikici, K. Idrissi, A. M. Baskurt, Institut National des Sciences Appliquйes de Lyon (France) . . . . . . . . . . . . . . . . . . . . . [6072-50] 4:20 pm: Asymmetrically informed data-hiding optimization of achievable rate for Laplacian host, J. E. Vila-Forcйn, O. J. Koval, S. V. Voloshynovskiy, E. Topak, T. Pun, Univ. de Genиve (Switzerland) [6072-51] 4:40 pm: Some theoretical aspects of watermarking detection, T. Furon, J. Josse, S. Le Squin, Institut National de Recherche en Informatique et en Automatique (France) . . . . . . . . . . . . . . . . . . . [6072-52] 5:00 pm: A framework for the design of good watermark identification codes, P. Moulin, R. Koetter, Univ. of Illinois at UrbanaChampaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-53] 5:20 pm: On the fundamental tradeoff between watermark detection performance and robustness against sensitivity analysis attacks, M. M. El Choubassi, P. Moulin, Univ. of Illinois at Urbana-Champaign[6072-16]
Thursday 19 January SESSION 13 Conv. Ctr. Room A5 . . . . . . . . . . . Thurs. 8:00 to 10:00 am Special Session: Benchmarking and Demonstration Session Chairs: Jana Dittmann, Otto-von-Guericke-Univ. Magdeburg (Germany); Benoоt B. Macq, Univ. Catholique de Louvain (Belgium) 8:00 am: Profiles for evaluation: the usage of audio WET, A. Lang, J. Dittmann, Otto-von-Guericke-Univ. Magdeburg (Germany) . . . . . [6072-54] 8:20 am: A benchmark assessment of the WAUC watermarking audio algorithm, D. Megнas Jimйnez, J. Herrera-Joancomarti, J. Serra i Ruiz, J. Minguillуn Alfonso, Univ. Oberta de Catalunya (Spain) . . . . . . . . [6072-55] 8:40 am: Transparency benchmarking on audio watermarks and steganography, C. Kraetzer, J. Dittmann, A. Lang, Otto-von-GuerickeUniv. Magdeburg (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-56] 9:00 am: Shape quality measurement for 3D watermarking schemes, P. Rondao-Alface, B. B. Macq, Univ. Catholique de Louvain (Belgium) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-57] 9:20 am: Reliability engineering approach to digital watermark evaluation, H. C. Kim, E. J. Delp III, O. Guitart, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-58] 9:40 am: New functionalities in watermark evaluation testbed (WET), O. Guitart, H. C. Kim, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . [6072-59] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:30 am SESSION 14 Conv. Ctr. Room A5 . . . . . . . . . . Thurs. 10:30 to 11:50 am Applications I Chair: Gordon W. Braudaway, IBM Corp. 10:30 am: Protection and governance of MPEG music player MAF contents using MPEG-21 IPMP tools, H. Hendry, M. Kim, Information and Communications Univ. (South Korea) . . . . . . . . . . . . . . . . . . [6072-60] 10:50 am: Watermarking of 3D objects based on 2D apparent contours, J. Bennour, J. Dugelay, Institut Eurйcom (France) . . . . [6072-61] 11:10 am: Quality assessment of watermarked 3D polygonal models, W. Funk, J. Prasiswa, Fraunhofer-Institut fьr Graphische Datenverarbeitung (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-62] 11:30 am: Reducing the processing time of the hierarchical watermark detector when applied to unmarked images, A. M. Alattar, O. M. Alattar, Digimarc Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-63] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:50 am to 1:20 pm SESSION 15 Conv. Ctr. Room A5 . . . . . . . . . . . . Thurs. 1:20 to 2:20 pm Applications II Chair: Min Wu, Univ. of Maryland/College Park 1:20 pm: Exploring QIM-based anti-collusion fingerprinting for multimedia, A. Swaminathan, S. He, M. Wu, Univ. of Maryland/College Park . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-64] 1:40 pm: Sheet music fingerprinting based on graphical representation, G. Kremser, M. Schmucker, Fraunhofer-Institut fьr Graphische Datenverarbeitung (Germany) . . . . . . . . . . . . . . . . . . [6072-66] 2:00 pm: A web-oriented and interactive buyer-seller watermarking protocol, F. Frattolillo, S. D'Onofrio, Univ. degli Studi del Sannio (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-67]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
53
Conference 6072 · Conv. Ctr. Room A5 SESSION 16 Conv. Ctr. Room A5 . . . . . . . . . . . . Thurs. 2:20 to 5:10 pm Embedding II Chair: Mauro Barni, Univ. degli Studi di Siena (Italy) 2:20 pm: Matrix embedding for large payloads, J. Fridrich, D. Soukal, Binghamton Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-68] 2:40 pm: Simple reversible watermarking schemes: further results, D. Coltuc, Univ. Valahia din Targoviste (Romania); J. Chassery, Institut National Polytechnique de Grenoble (France) . . . . . . . . . . . . . . . [6072-69] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm 3:30 pm: A new watermark detector for spread-spectrum based image watermarking using underdetermined independent component analysis framework, H. M. Malik, A. A. Khokhar, R. Ansari, Univ. of Illinois at Chicago . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-70] 3:50 pm: Optimal detector for an additive watermarking scheme based on human auditory system, M. Haddad, A. Gilloire, A. Le Guyader, France Tйlйcom (France); P. Duhamel, Lab. des signaux et systиmes (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-71] 4:10 pm: A hypothesis testing approach for achieving semi-fragility in multimedia authentication, C. Fei, Univ. of Toronto (Canada); D. Kundur, Texas A&M Univ.; R. Kwong, Univ. of Toronto (Canada) . . . . . . . . [6072-72] 4:30 pm: A DWT-based robust semi-blind image watermarking algorithm using two bands, E. Elbasi, A. M. Eskicioglu, The City Univ. of New York . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6072-73] 4:50 pm: Evaluating the visual quality of watermarked images, A. Shnayderman, A. M. Eskicioglu, The City Univ. of New York . . . . [6072-74]
54
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6073 · Conv. Ctr. Room B1
Tuesday-Thursday 17-19 January 2006 · Proceedings of SPIE Vol. 6073 Multimedia Content Analysis, Management, and Retrieval 2006
Conference Chairs: Edward Y. Chang, Univ. of California/Santa Barbara; Alan Hanjalic, Delft Univ. of Technology (Netherlands); Nicu Sebe, Univ. van Amsterdam (Netherlands) Program Committee: Kiyoharu Aizawa, The Univ. of Tokyo (Japan); Aya Aner-Wolf, GenTech Corp. (Israel); Noboru Babaguchi, Osaka Univ. (Japan); Nozha Boujemaa, INRIA Rocquencourt (France); Arbee L. P. Chen, National Chengchi Univ. (Taiwan); Tsuhan Chen, Carnegie Mellon Univ.; TatSeng Chua, National Univ. of Singapore (Singapore); Ajay Divakaran, Mitsubishi Electric Research Labs.; Chitra Dorai, IBM Thomas J. Watson Research Ctr.; Arun Hampapur, IBM Thomas J. Watson Research Ctr.; Alexander G. Hauptmann, Carnegie Mellon Univ.; Alejandro Jaimes, Fuji Xerox Co., Ltd. (Japan); Mohan S. Kankanhalli, National Univ. of Singapore (Singapore); John R. Kender, Columbia Univ.; Anil C. Kokaram, The Univ. of Dublin, Trinity College (Ireland); Michael S. Lew, Leiden Institute of Advanced Computer Science (Netherlands); Chung-Sheng Li, IBM Corp.; Rainer W. Lienhart, Univ. of Augsburg (Germany); Wei-Ying Ma, Microsoft Research Asia (China); Bernard Merialdo, Institut Eurйcom (France); Kadir A. Peker, Mitsubishi Electric Research Labs.; Silvia Pfeiffer, Commonwealth Scientific & Industrial Research Organisation (Australia); Alan F. Smeaton, Dublin City Univ. (Ireland); John R. Smith, IBM Thomas J. Watson Research Ctr.; Hari Sundaram, Arizona State Univ.; Ahmet M. Tekalp, Univ. of Rochester; Qi Tian, The Univ. of Texas at San Antonio; Svetha Venkatesh, Curtin Univ. of Technology (Australia); Stephen T. C. Wong, Harvard Medical School; Marcel Worring, Univ. van Amsterdam (Netherlands); Aidong Zhang, SUNY/Univ. at Buffalo
Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. A tree-based paradigm for content-based video retrieval and management, H. Fang, Univ. of Bradford (United Kingdom); Y. Yin, Chongqing Univ. (China); J. Jiang, Univ. of Bradford (United Kingdom) [6073-31] Tangible interactive system for document browsing and visualization of multimedia data, Y. B. Rytsar, S. V. Voloshynovskiy, O. J. Koval, F. Deguillaume, E. Topak, Univ de Genиve (Switzerland); S. Startchik, Anteleon Imaging (Switzerland); T. Pun, Univ. de Genиve (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-32] Semantic segmentation of video collections using boosted random fields, B. Janvier, E. Bruno, S. Marchand-Maillet, T. Pun, Univ. de Genиve (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-33] Annotating 3D contents with MPEG-7 for reuse purposes, I. M. Bilasco, J. Gensel, M. Villanova-Oliver, H. Martin, Institut d'Informatique et Mathйmatiques Appliquйes de Grenoble (France) . . . . . . . . [6073-34] Multimedia for art retrieval (M4ART), E. L. van den Broek, Vrije Univ. Amsterdam (Netherlands); T. Kok, T. E. Schouten, E. Hoenkamp, Radboud Univ. Nijmegen (Netherlands) . . . . . . . . . . . . . . . . . . [6073-35]
Application of image visual characterization and soft feature selection in content-based image retrieval, K. Jarrah, I. Lee, Ryerson Univ. (Canada); M. J. Kyan, The Univ. of Sydney (Australia); L. Guan, Ryerson Univ. (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-36] Video shot retrieval using a kernel derived from a continuous HMM, A. Velivelli, T. S. Huang, Univ. of Illinois at Urbana-Champaign; A. G. Hauptmann, Carnegie Mellon Univ. . . . . . . . . . . . . . . . . . . . . . [6073-37] Moving camera moving object segmentation in an MPEG-2 compressed video sequence, J. Wang, Wanye State Univ.; N. Patel, W. Grosky, Univ. of Michigan/Dearborn . . . . . . . . . . . . . . . . . . [6073-38] Visual object categorization with indefinite kernels in discriminant analysis framework, S. Kosinov, S. Marchand-Maillet, Univ. de Genиve (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-39] Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details. SESSION 1 Conv. Ctr. Room B1 . . . . . . . . . . . . Wed. 9:30 to 10:30 am Video Analysis I Chair: Alan F. Smeaton, Dublin City Univ. (Ireland) 9:30 am: Blind summarization: content adaptive video summarization using time-series analysis, A. Divakaran, R. Radhakrishnan, K. A. Peker, Mitsubishi Electric Research Labs. . . . . . . . . . . . . . . . . . . . . . . . . [6073-01] 9:50 am: Multilevel analysis of sports video sequences, J. Han, D. Farin, P. H. N. de With, Technische Univ. Eindhoven (Netherlands) . . . . [6073-02] 10:10 am: Automated editing of medical training video via content analysis, A. C. Kokaram, The Univ. of Dublin, Trinity College (Ireland); K. Andrews, Univ. of the West Indies (Trinidad and Tobago); D. Ring, The Univ. of Dublin, Trinity College (Ireland); C. Lee, Royal College of Surgeons in Ireland (Ireland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-03] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 10:50 am
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
55
Conference 6073 · Conv. Ctr. Room B1
SESSION 2 Conv. Ctr. Room B1 . . . . . . . . Wed. 10:50 am to 12:10 pm Audio and Video Retrieval Chair: Anil C. Kokaram, The Univ. of Dublin, Trinity College (Ireland) 10:50 am: Statistical model and error analysis of a proposed audio fingerprinting algorithm, E. P. McCarthy, F. Balado, N. J. Hurley, G. C. M. Silvestre, Univ. College Dublin (Ireland) . . . . . . . . . . . . . . . . . . . . [6073-04] 11:10 am: An application of weighted transducers to music information retrieval, D. Basaldella, N. Orio, Univ. degli Studi di Padova (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-05] 11:30 am: Video scene retrieval with symbol sequence based on integrated audio and visual features, K. Morisawa, N. Nitta, N. Babaguchi, Osaka Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-06] 11:50 am: Fнschlбr-DiamondTouch: collaborative video searching on a table, A. F. Smeaton, H. Lee, C. Foley, S. McGivney, C. Gurrin, Dublin City Univ. (Ireland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-07] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:40 pm SESSION 3 Conv. Ctr. Room B1 . . . . . . . . . . . . . Wed. 1:40 to 3:00 pm Image Retrieval Chair: Nicu Sebe, Univ. van Amsterdam (Netherlands) 1:40 pm: Mind the gap: another look at the problem of the semantic gap in image retrieval, J. S. Hare, P. H. Lewis, Univ. of Southampton (United Kingdom); P. Enser, C. Sandom, Univ. of Brighton (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-08] 2:00 pm: Evaluation of strategies for multiple sphere queries with local image descriptors, N. Bouteldja, Conservatoire National des Arts et Metiers/CEDRIC (France); V. Gouet-Brunet, M. Scholl, Conservatoire National des Arts et Mйtiers/CEDRIC (France) . . . . . . . . . . . . . . . [6073-09] 2:20 pm: PARIS: a MPEG-7 spatial and temporal referenced personal photograph library, P. Kuo, M. Ito, T. Aoki, H. Yasuda, The Univ. of Tokyo (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-10] 2:40 pm: 2+2=5: painting by numbers, C. C. Venters, Univ. of Manchester (United Kingdom); R. J. Hartley, Manchester Metropolitan Univ. (United Kingdom); W. T. Hewitt, Univ. of Manchester (United Kingdom) . [6073-11] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:20 pm SESSION 4 Conv. Ctr. Room B1 . . . . . . . . . . . . . Wed. 3:20 to 4:20 pm Applications I Chair: Alan Hanjalic, Technische Univ. Delft (Netherlands) 3:20 pm: Structuring continuous video recordings of everyday life using time-constrained clustering, W. Lin, A. G. Hauptmann, Carnegie Mellon Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-12] 3:40 pm: Practical life log video indexing based on content and context, D. Tancharoen, T. Yamasaki, K. Aizawa, The Univ. of Tokyo (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-13] 4:00 pm: Multimedia for mobile users: image-enhanced navigation, S. Gautam, G. Sarkis, E. Tjandranegara, E. Zelkowitz, Y. Lu, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-14] Panel Discussion . . . . . . . . . . . . . . . . . . . 4:20 to 5:20 pm Chair: Rainer W. Lienhart, Univ. Augsburg (Germany)
Thursday 19 January SESSION 5 Conv. Ctr. Room B1 . . . . . . . . . . . Thurs. 9:00 to 10:00 am Image Classification Chair: Michael G. Christel, Carnegie Mellon Univ. 9:00 am: Semantic classification of business images, B. Erol, J. J. Hull, Ricoh Innovations, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-15] 9:20 am: Region labeling using a point-based coherence criterion, H. Houissa, N. Boujemaa, INRIA Rocquencourt (France) . . . . . . . . . [6073-16] 9:40 am: BlobContours: adapting Blobworld for supervised color- and texture-based image segmentation, T. Vogel, D. N. Quyen, J. Dittmann, Otto-von-Guericke-Univ. Magdeburg (Germany) . . . . . . . . . . . . . [6073-17] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:00 to 10:20 am SESSION 6 Conv. Ctr. Room B1 . . . . . . . Thurs. 10:20 am to 12:20 pm Special Session: Evaluating Video Summarization, Browsing, and Retrieval Techniques Chair: Ajay Divakaran, Mitsubishi Electric Research Labs. 10:20 am: Subjective assessment of consumer video summarization (Invited Paper), C. Forlines, K. A. Peker, A. Divakaran, Mitsubishi Electric Research Labs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-18] 10:50 am: Evaluation of video summarization systems (Invited Paper), C. M. Taskiran, Motorola, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-19] 11:20 am: Subjective evaluation criterion for selecting affective features and modeling highlights (Invited Paper), L. Xing, H. Yu, Q. Huang, Graduate School of the Chinese Academy of Sciences (China); Q. Ye, Chinese Academy of Sciences (China); A. Divakaran, Mitsubishi Electric Research Labs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-20] 11:50 am: Evaluation and user studies with respect to video summarization and browsing (Invited Paper), M. G. Christel, Carnegie Mellon Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-21] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:20 to 2:00 pm SESSION 7 Conv. Ctr. Room B1 . . . . . . . . . . . . Thurs. 2:00 to 3:00 pm Feature Extraction Chair: Edward Y. Chang, Univ. of California/Santa Barbara 2:00 pm: Semantic feature extraction with multidimensional hidden Markov model, J. Jiten, B. Merialdo, B. Huet, Eurйcom Institute (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-22] 2:20 pm: Rotation and translation invariant feature extraction using angular projection in frequency domain, B. Lee, M. Kim, Information and Communications Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . [6073-23] 2:40 pm: Invariant region descriptors for robust shot segmentation, A. Arasanathan, N. Canagarajah, Univ. of Bristol (United Kingdom) . [6073-24] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:20 pm
56
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6073 · Conv. Ctr. Room B1 SESSION 8 Conv. Ctr. Room B1 . . . . . . . . . . . . Thurs. 3:20 to 4:20 pm Video Analysis II Chair: Rainer W. Lienhart, Univ. Augsburg (Germany) 3:20 pm: A video processing method for convenient mobile reading of printed barcodes with camera phones, C. H. Bдckstrцm, C. Sцdergеrd, VTT Information Technology (Finland); S. Udd, UPC Consulting Ltd. (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-25] 3:40 pm: Flexible surveillance system architecture for prototyping video content analysis algorithms, R. Wijnhoven, Bosch Security Systems B.V. (Netherlands); E. G. T. Jaspers, P. H. N. de With, LogicaCMG (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6073-26] 4:00 pm: Motion-based parsing for video from observational psychology, A. C. Kokaram, E. Doyle, D. Lennon, L. Joyeux, R. Fuller, The Univ. of Dublin, Trinity College (Ireland) . . . . . . . . . . . . . . . . . . . . [6073-27] Session Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4:20 to 4:30 pm SESSION 9 Conv. Ctr. Room B1 . . . . . . . . . . . . Thurs. 4:30 to 5:10 pm Applications II Chair: Berna Erol, Ricoh Innovations, Inc. 4:30 pm: Occlusion costing for multimedia object layout in a constrained window, S. Widdowson, Hewlett Packard Labs. . . [6073-28] 4:50 pm: Using CART to segment road images, R. L. Davies, Intel Corp.; R. W. Lienhart, Univ. Augsburg (Germany) . . . . . . . . . . . . . . . . . . [6073-30]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
57
Conference 6074 · Conv. Ctr. Room B4
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6074 Multimedia on Mobile Devices II
Conference Chairs: Reiner Creutzburg, Fachhochschule Brandenburg (Germany); Jarmo H. Takala, Tampere Univ. of Technology (Finland); Chang Wen Chen, Florida Institute of Technology Program Committee: David Akopian, The Univ. of Texas at San Antonio; Alan Chalmers, Univ. of Bristol (United Kingdom); Surendar Chandra, Univ. of Notre Dame; Kenneth J. Crisler, Motorola Labs.; David S. Doermann, Univ. of Maryland/College Park; Uwe Dummann, Siemens AG (Germany); Elizabeth Dykstra-Erickson, Kinoma, Inc.; Zhihai He, Univ. of Missouri/Columbia; Xin Li, West Virginia Univ.; Sethuraman Panchanathan, Arizona State Univ.; Kari A. Pulli, Nokia; Matthias Rauterberg, Technische Univ. Eindhoven (Netherlands); Phillip A. Regalia, Institut National des Tйlйcommunications (France); Haitao Zheng, Univ. of California/Santa Barbara
Monday 16 January SESSION 1 Conv. Ctr. Room B4 . . . . . . . . . . . . Mon. 9:30 to 10:00 am Invited Paper I 9:30 am: New APIs for mobile graphics (Invited Paper, Presentation Only), K. A. Pulli, Nokia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-01] SESSION 2 Conv. Ctr. Room B4 . . . . . . . . . . . Mon. 10:00 to 10:20 am Multimedia Coding I Chair: Reiner Creutzburg, Fachhochschule Brandenburg (Germany) 10:00 am: A novel fast inter-prediction mode decision for H.264/AVC, Y. Guo, H. Li, S. Pei, Univ. of Science and Technology of China (China); C. W. Chen, Florida Institute of Technology . . . . . . . . . . . . . . . . . . . [6074-02] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am SESSION 3 Conv. Ctr. Room B4 . . . . . . . . Mon. 10:50 am to 12:10 pm Multimedia Coding II Chair: Chang Wen Chen, Florida Institute of Technology 10:50 am: Concealment driven bit rate reduction using the H.264 video coding standard, S. T. Beesley, A. J. Armstrong, C. Grecos, D. J. Parish, Loughborough Univ. (United Kingdom) . . . . . . . . . . . . . . . . . . . . [6074-03] 11:10 am: Image embedded coding with edge preservation based on local variance analysis for mobile applications, G. Luo, D. Osypiw, Buckinghamshire Chilterns Univ. College (United Kingdom) . . . . [6074-04] 11:30 am: Image coding using adaptive resizing in the block-DCT domain, J. J. Koivusaari, J. H. Takala, M. Gabbouj, Tampere Univ. of Technology (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-05] 11:50 am: Spatial scalability of multiple ROIs in scalable video coding, T. M. Bae, T. C. Thang, D. Y. Kim, Y. M. Ro, Information and Communications Univ. (South Korea); J. Kim, Electronics and Telecommunications Research Institute (South Korea) . . . . . . . . [6074-06] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 2:00 pm SESSION 4 Conv. Ctr. Room B4 . . . . . . . . . . . . . Mon. 2:00 to 2:30 pm Invited Paper II 2:00 pm: Implementing energy efficient embedded multimedia (Invited Paper), O. J. Silvйn, Univ. of Oulu (Finland) . . . . . . . . . . . . . . . . . [6074-07]
SESSION 5 Conv. Ctr. Room B4 . . . . . . . . . . . . . Mon. 2:30 to 3:10 pm Mobile Multimedia Retrieval and Classification Chair: Stefan Edlich, Technische Fachhochschule Berlin (Germany) 2:30 pm: A study of low-complexity tools for semantic classification of mobile images and video, A. Mariappan, M. Igarta, Purdue Univ.; C. M. Taskiran, B. Gandhi, Motorola, Inc.; E. J. Delp III, Purdue Univ. . . [6074-08] 2:50 pm: Audio-based queries for video retrieval over Java enabled mobile devices, I. Ahmad, Nokia Corp. (Finland); F. A. Cheikh, S. Kiranyaz, M. Gabbouj, Tampere Univ. of Technology (Finland) . . . . . . . . . . [6074-09] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:30 pm SESSION 6 Conv. Ctr. Room B4 . . . . . . . . . . . . . Mon. 3:30 to 5:10 pm Processors for Multimedia Chair: Jarmo H. Takala, Tampere Univ. of Technology (Finland) 3:30 pm: Parallel implementation of MPEG-2 video decoder, A. Sarkar, North Carolina State Univ.; K. Saha, S. Maiti, STMicroelectronics Pvt. Ltd. (India) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-11] 3:50 pm: Software-based geometry operations for 3D computer graphics, M. Sima, Univ. of Victoria (Canada); D. Iancu, J. C. Glossner, Sandbridge Technologies, Inc.; M. J. Schulte, S. Mamidi, Univ. of Wisconsin/Madison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-12] 4:10 pm: MVSP: multithreaded VLIW stream processor, S. Sardashti, H. R. Ghasemi, O. Fatemi, Univ. of Tehran (Iran) . . . . . . . . . . . . . . . . [6074-13] 4:30 pm: System-on-chip architecture with media DSP and RISC core for media application, P. Liu, W. Wang, Q. Yao, Zhejiang Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-14] 4:50 pm: IIA: a novel method to optimize media instruction set of embedded processor, K. Chen, Q. Yao, W. Wang, P. Liu, Zhejiang Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-15] Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 7 Conv. Ctr. Room B4 . . . . . . . . . . . . Tues. 9:30 to 10:00 am Invited Paper III 9:30 am: Multimedia services for next-generation mobile networks (Invited Paper), S. J. Wee, D. Penkler, Hewlett-Packard Labs. . . [6074-17]
58
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6074 · Conv. Ctr. Room B4
SESSION 8 Conv. Ctr. Room B4 . . . . . . . . . . . Tues. 10:00 to 10:20 am Multimedia Applications I Chair: Reiner Creutzburg, Fachhochschule Brandenburg (Germany) 10:00 am: Wireless steganography, S. S. Agaian, D. Akopian, S. D'Souza, The Univ. of Texas at San Antonio . . . . . . . . . . . . . . . . . . . . . . . . [6074-18] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am SESSION 9 Conv. Ctr. Room B4 . . . . . . . . Tues. 10:50 am to 12:10 pm Multimedia Applications II Chair: Zhihai He, Univ. of Missouri/Columbia 10:50 am: Image processing for navigation on a mobile embedded platform, H. Loose, C. Lemke, C. Papazov, Brandenburg Univ. of Applied Sciences (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-19] 11:10 am: Image processing for navigation on a mobile embedded platform, T. Preuss, L. Gentsch, M. Rambow, Brandenburg Univ. of Applied Sciences (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-20] 11:30 am: The future is `ambient', A. R. Lugmayr, Tampere Univ. of Technology (Finland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-21] 11:50 am: Embedded video surveillance system for vehicle over WLAN and CDMA1X, L. Ming, Wuhan Univ. (China) . . . . . . . . . . . . . . . . [6074-22] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:10 to 1:40 pm SESSION 10 Conv. Ctr. Room B4 . . . . . . . . . . . . . Tues. 1:40 to 3:00 pm Multimedia Data Management Chair: David Akopian, The Univ. of Texas at San Antonio 1:40 pm: Performance analysis of MPEG-21 technologies on mobile devices, S. De Zutter, F. De Keukelaere, C. Poppe, R. Van de Walle, Univ. Gent (Belgium) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-23] 2:00 pm: TV-anytime and MPEG-21 DIA based ubiquitous content mobility prototype system for multi-users, M. Kim, C. Yang, J. Lim, M. Kim, Information and Communications Univ. (South Korea); S. Park, K. Kim, SAMSUNG Electronics Co., Ltd. (South Korea) . . . . . . . . . . [6074-24] 2:20 pm: Multimedia-based construction management and supervision on mobile JTWI devices and the EDF-portal, S. Edlich, Technische Fachhochschule Berlin (Germany); R. Strauch, edv plan GmbH; L. Visengeriyeva, Fachhochschule Brandenburg; D. Reeck, edv plan GmbH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-25] 2:40 pm: A mobile phone-based context-aware video management application, J. Lahti, M. Palola, J. Korva, VTT Technical Research Ctr. of Finland (Finland); U. Westermann, P. Pietarila, VTT Elektroniikka (Finland); K. Pentikousis, VTT Technical Research Ctr. of Finland (Finland) . [6074-26] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm
SESSION 11 Conv. Ctr. Room B4 . . . . . . . . . . . . . Tues. 3:30 to 5:10 pm HCI Issues for Mobile Devices Chair: Jarmo H. Takala, Tampere Univ. of Technology (Finland) 3:30 pm: MIKE's PET: a participant-based experiment tracking tool for HCI practitioners using mobile devices, D. Mohamedally, City Univ. London (United Kingdom); S. Edlich, Technische Fachhochschule Berlin (Germany); E. Klaus, Fachhochschule Brandenburg (Germany); P. Zaphiris, H. Petrie, City Univ. London (United Kingdom) . . . . . . . . . . . . . . . [6074-27] 3:50 pm: Maintenance support: case study for a multimodal mobile interface, G. Fuchs, D. Reichart, H. Schumann, P. Forbrig, Univ. Rostock (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-28] 4:10 pm: Breaking the news on mobile TV: user requirements of a popular mobile content, H. O. Knoche, Univ. College London; A. Sasse, Univ. College London (United Kingdom) . . . . . . . . . . . . . . . . . . . . [6074-29] 4:30 pm: Multimodal audio guide for museums and exhibitions, S. Gebbensleben, J. Dittmann, Otto-von-Guericke Univ. (Germany) [6074-30] 4:50 pm: Human sound detection on experience movies, S. Shimura, Y. Hirano, S. Kajita, K. Mase, Nagoya Univ. (Japan) . . . . . . . . . . . . . [6074-31] Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . .5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. A FGS coding method based on LC multiwavelet transform, W. Liu, South China Normal Univ. (China); Z. Ma, Zhongshan Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-34] A context-aware video display scheme for mobile devices, K. Seo, C. Kim, Information and Communications Univ. (South Korea) [6074-35] Verification of WIPI-based T-DMB platform for interactive mobile multimedia services, B. Bae, W. Kim, J. Yun, C. Ahn, Electronics and Telecommunications Research Institute (South Korea); K. Sohng, Kyungpook National Univ. (South Korea) . . . . . . . . . . . . . . . . . [6074-36] New TPEG applications based on digital multimedia broadcasting, Y. Jeong, S. Cho, G. Kim, C. Ahn, S. Lee, Electronics and Telecommunications Research Institute (South Korea); W. Kim, Chungnam National Univ. (South Korea) . . . . . . . . . . . . . . . . . [6074-37] AROMA: augmented reality on mobile devices API (Java), S. Edlich, Technische Fachhochschule Berlin (Germany); H. Hцrning, R. Hцrning, Biting Bit (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-39] An effective method and its implementation for splicing in terrestrial DMB, Y. Lee, J. Lee, G. Lee, C. Ahn, S. Lee, Electronics and Telecommunications Research Institute (South Korea); N. Kim, Chungbuk National Univ. (South Korea) . . . . . . . . . . . . . . . . . . [6074-40] Media digital signal processor core design for multimedia application, P. Liu, G. Yu, W. Cai, Q. Yao, Zhejiang Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6074-42]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
59
Conference 6075 · Conv. Ctr. Room A7
Tuesday-Thursday 17-19 January 2006 · Proceedings of SPIE Vol. 6075 Optical Security and Counterfeit Deterrence Techniques VI
Conference Chair: Rudolf L. van Renesse, VanRenesse Consulting (Netherlands) Program Committee: Sara E. Church, Bank of Canada (Canada); James M. Jonza, 3M Co.; Malcolm R. M. Knight, De La Rue International Ltd. (United Kingdom); Ian M. Lancaster, Reconnaissance International (United Kingdom) and International Hologram Manufacturers Association (United Kingdom); Hiroyuki Matsumoto, NHK Spring Co., Ltd. (Japan); Roger W. Phillips, JDSU-Flex Products Group; Elisabeth Schulz, European Central Bank (Germany); Sybrand Spannenburg, Joh. Enschedй Security Printing B.V. (Netherlands); Wayne R. Tompkin, OVD Kinegram Corp. (Switzerland); Jan van den Berg, Sdu Identification (Netherlands)
Tuesday 17 January
SESSION 1
Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Use of metameric filters for future interference security images structures, B. Baloukas, L. Martinu, Йcole Polytechnique de Montrйal (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-42] Recording the optical identification marks on CDs, A. A. Kryuchyn, Institute for Information Recording (Ukraine); S. A. Kostyukevych, V. Lashkaryov Institute of Semiconductor Physics (Ukraine); V. V. Petrov, S. M. Shanoylo, Institute for Information Recording (Ukraine) . . . [6075-40] Concealed images in design of security devices and methods of their authentication, V. I. Girnyk, Optronics, Ltd. (Ukraine); S. O. Kostyukevich, Optronics, Ltd. (Ukraine) and Institute of Semiconductor Physics (Ukraine); E. V. Braginets, Optronics, Ltd. (Ukraine) and National Taras Shevchenko Univ. of Kyiv (Ukraine); A. Soroka, National Taras Shevchenko Univ. of Kyiv (Ukraine) . . . . . . . . . . . . . . . . . . . . . [6075-41] Wednesday 18 January Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details.
Conv. Ctr. Room A7 . . . . . . . . . . . . Wed. 9:30 to 10:30 am Currency I Chair: Sara E. Church, Bank of Canada (Canada) 9:30 am: Feed back from the public for better banknote design, H. A. M. de Heij, De Nederlandsche Bank (Netherlands) . . . . . . . . . . . . . . [6075-01] 9:50 am: Robust and reliable banknote authentification and print flaw detection with opto-acoustical sensor fusion methods, V. Lohweg, Fachhochschule Lippe und Hцxter (Germany); J. G. Schaede, KBA-GIORI S.A. (Switzerland); T. Tьrke, KBA-Bielefeld (Germany) . . . . . . . . . [6075-02] 10:10 am: The implication of direct laser engraved intaglio plates on bank note security, H. Deinhammer, Цsterrreichische Banknoten- und Sicherheitsdruck (Austria) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-03] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am SESSION 2 Conv. Ctr. Room A7 . . . . . . . . Wed. 11:00 am to 12:00 pm Currency II Chair: Elisabeth Schulz, European Central Bank (Germany) 11:00 am: The circulation simulator method for evaluating bank note and optical feature durability, W. J. Bartz, Crane & Co., Inc. . . [6075-04] 11:20 am: Visual and optical evaluation of bank notes in circulation, S. E. Church, M. P. Lacelle, T. Garanzotis, Bank of Canada (Canada) . [6075-05] 11:40 am: The mechanisms of human recognition as a guideline for security feature development, J. G. Schaede, KBA-GIORI S.A. (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-06] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm SESSION 3 Conv. Ctr. Room A7 . . . . . . . . . . . . . Wed. 1:30 to 3:10 pm Security Image Technology I Chair: Sybrand Spannenburg, Joh. Enschedй Security Printing B.V. (Netherlands) 1:30 pm: Qualification of security printing features, S. J. Simske, HewlettPackard Labs; J. S. Aronoff, Hewlett-Packard Labs.; J. Arnabat, HewlettPackard Co. (Spain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-07] 1:50 pm: Potentiality of holographic technique in fragile watermarking (Invited Paper), G. Schirripa Spagnolo, Univ. degli Studi di Roma Tre (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-08] 2:10 pm: Secure graphical data storage by full-spectrum image coding, H. Oltmans, Consultant (Netherlands) . . . . . . . . . . . . . . . . . . . . . [6075-09] 2:30 pm: Sub-pixel analysis to support graphic security after scanning at low resolution, R. A. Cordery, Pitney Bowes; S. K. Decker, Digimarc Corp.; B. Haas, Pitney Bowes; H. Gou, Univ. of Maryland/College Park . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-10] 2:50 pm: New applications of modulated digital images in document security, R. Lee, P. W. Leech, L. D. McCarthy, G. F. Swiegers, Commonwealth Scientific & Industrial Research Organisation (Australia) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-11] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm
60
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6075 · Conv. Ctr. Room A7
SESSION 4 Conv. Ctr. Room A7 . . . . . . . . . . . . . Wed. 3:40 to 5:20 pm Security Image Technology II Chair: James M. Jonza, 3M Co. 3:40 pm: Black fluorescent ink and applications, J. Auslander, R. A. Cordery, Pitney Bowes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-37] 4:00 pm: The role of optics in secure credentials, T. L. Lichtenstein, LaserCard Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-13] 4:20 pm: Practical use of lens structures in ID documents, J. van den Berg, SDU Identification (Netherlands) . . . . . . . . . . . . . . . . . . . . . [6075-14] 4:40 pm: Three-dimensional floating images as overt security features, D. S. Dunn, T. L. Potts, L. E. Lorimor, J. M. Jonza, R. L. Smithson, S. P. Maki, 3M Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-15] 5:00 pm: Development of the random-retardation-encoding anticounterfeiting technology, W. Huang, C. Tsai, T. Chen, M. Kuan, C. Wen, Industrial Technology Research Institute (Taiwan) . . . . . . . . . . . . [6075-16] Thursday 19 January SESSION 5 Conv. Ctr. Room A7 . . . . . . . . . . . Thurs. 8:10 to 10:10 am Security Ink Technology Chair: Roger W. Phillips, JDSU-Flex Products Group 8:10 am: Advanced verification methods for OVI(r) security ink, P. G. Coombs, Flex Products/JDS Uniphase Corp.; S. F. McCaffery, JDS Uniphase Corp.; T. Markantes, Flex Products/JDS Uniphase Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-17] 8:30 am: Overt security features through digital printing, R. A. Einhorn, M. J. Hampden-Smith, S. Haubrich, J. Shah, R. Bhatia, N. Hardman, R. Kornbrekke, Cabot Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-18] 8:50 am: Novel particulate production processes to create unique security materials, M. J. Hampden-Smith, T. Kodas, S. Haubrich, M. Oljaca, R. A. Einhorn, Cabot Corp. . . . . . . . . . . . . . . . . . . . . . . . . [6075-19] 9:10 am: Combining overt and covert anti-counterfeiting technologies for securities, T. Uematsu, National Printing Bureau of Japan (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-20] 9:30 am: Bacteriorhodopsin-based multilevel optical security features, N. A. Hampp, M. Neebe, I. Yang, Philipps-Univ. Marburg (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-21] 9:50 am: Optical security in ink: an industry standard that continues to evolve, M. Schmid, SICPA SA (Switzerland) . . . . . . . . . . . . . . . . [6075-22] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am SESSION 6 Conv. Ctr. Room A7 . . . . . . . Thurs. 10:40 am to 12:00 pm Optically Variable Devices I Chair: Wayne R. Tompkin, OVD Kinegram AG (Switzerland) 10:40 am: Current use and efficacy of optical security devices, I. M. Lancaster, Reconnaissance International Ltd. (United Kingdom) and International Hologram Manufacturing Association (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-23] 11:00 am: The security enhancement of diffractive optically variable image devices, A. Argoitia, R. W. Phillips, Flex Products/JDSU . [6075-24] 11:20 am: The Aztec structure: an improved replicable security device, J. J. Cowan, Aztec Systems, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . [6075-25] 11:40 am: Combination of optically variable diffractive and multilayer interference structures: a new class of security devices, V. I. Girnyk, Optronics, Ltd. (Ukraine); R. W. Phillips, JDSU-Flex Products Group; E. V. Braginets, Optronics, Ltd. (Ukraine) and National Kiev Taras Schevchenko Univ. (Ukraine) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-26] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm
SESSION 7 Conv. Ctr. Room A7 . . . . . . . . . . . . Thurs. 1:30 to 2:30 pm Optically Variable Devices II Chair: Ian M. Lancaster, Reconnaissance International Ltd. (United Kingdom) 1:30 pm: Novel optically variable color devices, M. Stalder, F. Seils, Rolic Technologies Ltd. (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-27] 1:50 pm: Diffractive Moirй features for optically variable devices, A. Schilling, W. R. Tompkin, R. Staub, OVD Kinegram AG (Switzerland); R. D. Hersch, S. Chosson, I. Amidror, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-28] 2:10 pm: Combination of e-beam and optical holography on microand macro-levels of OVD, E. V. Braginets, V. I. Girnyk, Optronics, Ltd. (Ukraine) and National Taras Shevchenko Univ. (Ukraine); B. Holmes, De La Rue International Ltd. (United Kingdom) . . . . . . . . . . . . . . . . . . . [6075-29] SESSION 8 Conv. Ctr. Room A7 . . . . . . . . . . . . Thurs. 2:30 to 4:20 pm Authentication, Identification, and Biometrics I Chair: Malcolm R. M. Knight, De La Rue International Ltd. (United Kingdom) 2:30 pm: Choosing the correct forensic marker(s) in currency, document, and product protection, J. J. Plimmer, Product & Image Security Foundation (United Kingdom) . . . . . . . . . . . . . . . . . . . . . [6075-30] 2:50 pm: Optically variable threads and polarization effects, F. Kretschmar, Louisenthal GmbH (Germany) . . . . . . . . . . . . . . . . . [6075-38] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm 3:40 pm: Life recognition based on color variations in fingerprint images of live and artificial fingers, K. Tai, M. Kurita, I. Fujieda, Ritsumeikan Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-32] 4:00 pm: RFID identity theft and countermeasures, A. Herrigel, UpGreat AG (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-33] SESSION 9 Conv. Ctr. Room A7 . . . . . . . . . . . . Thurs. 4:20 to 5:20 pm Authentication, Identification, and Biometrics II Chair: Rudolf L. van Renesse, VanRenesse Consulting (Netherlands) 4:20 pm: On the use of mobile imaging devices for the validation of first and second line security features, T. F. Rodriguez, M. Weaver III, Digimarc Corp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-34] 4:40 pm: Facets of color laser marking in high secure ID documents, F. Kappe, M. Schumacher, ORGA Systems enabling services GmbH (Germany); K. Schдfer, M. Hillebrand, M. Hennemeyer-Schwenkner, D. Fischer, Orga Systems GmbH (Germany) . . . . . . . . . . . . . . . . . . . [6075-35] 5:00 pm: Protection of data carriers using secure optical codes, J. A. Peters, A. Schilling, R. Staub, W. R. Tompkin, OVD Kinegram AG (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6075-36] Standby Oral Presentation 5:40 am: Public education by Central Banks on the Internet, R. L. van Renesse, VanRenesse Consulting (Netherlands) . . . . . . . . . . . . . [6075-39]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
61
Conference 6076 · Conv. Ctr. Room A7
Monday-Tuesday 16-17 January 2006 · Proceedings of SPIE Vol. 6076 Digital Publishing
Conference Chairs: Jan P. Allebach, Purdue Univ.; Hui Chao, Hewlett-Packard Co. Program Committee: Kathrin Berkner, Ricoh Innovations, Inc.; Charles A. Bouman, Purdue Univ.; David F. Brailsford, Univ. of Nottingham (United Kingdom); Richard Furuta, Texas A&M Univ.; Steven J. Harrington, Xerox Corp.; Yuukou Horita, Toyama Univ. (Japan); Charles Jacobs, Microsoft Corp.; Dhiraj Kacker, Caslon & Co.; John Lumley, Hewlett-Packard Ltd. (United Kingdom); Lisa Purvis, Xerox Corp.; Fernando Vega, Univ. de Puerto Rico Mayagьez; Fabio Vitali, Univ. degli Studi di Bologna (Italy)
Monday 16 January SESSION 1 Conv. Ctr. Room A7 . . . . . . . . . Mon. 9:30 am to 12:00 pm New Publishing Methods Chair: Steven J. Harrington, Xerox Corp. 9:30 am: E-books and the challenge of preservation (Invited Paper), F. Romano, Rochester Institute of Technology . . . . . . . . . . . . . . . . . [6076-01] 10:10 am: User centered design of the digital book: why looking backward can help us move forward, J. C. Wallis, Univ. of California/Los Angeles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-02] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am 11:00 am: Interactive publications: creation and usage, G. R. Thoma, G. Ford, M. Chung, K. Vasudevan, S. K. Antani, National Library of Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-03] 11:20 am: Personalized direct marketing using digital publishing, L. K. Cheeniyil, J. K. Prabhakaran, Hewlett-Packard Co. (India) . . . . . [6076-05] 11:40 am: Automated campaign system, G. L. Vondran, H. Chao, X. Lin, P. Joshi, D. Beyer, C. B. Atkins, P. Obrador, Hewlett-Packard Co. [6076-06] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:30 pm SESSION 2 Conv. Ctr. Room A7 . . . . . . . . . . . . . Mon. 1:30 to 3:10 pm Document Structure and Style Chair: John Lumley, Hewlett-Packard Ltd. (United Kingdom) 1:30 pm: Is the nature of a document changing (Invited Paper), J. C. King, S. Towers, Adobe Systems, Inc. . . . . . . . . . . . . . . . . . . . . . [6076-07] 2:10 pm: Expression of document structure in automatic layout, S. J. Harrington, Xerox Corp.; R. Price Jones, J. F. Naveda, Rochester Institute of Technology; N. Thakkar, IBM Corp. . . . . . . . . . . . . . . . . . . . . . [6076-08] 2:30 pm: Evaluating interface aesthetics: measure of symmetry, H. Balinsky, Hewlett-Packard Ltd. (United Kingdom) . . . . . . . . . . . . [6076-09] 2:50 pm: Automatic color scheme picker for document templates based on image analysis and dual problem, P. Obrador, HewlettPackard Labs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-10] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm
SESSION 3 Conv. Ctr. Room A7 . . . . . . . . . . . . . Mon. 3:40 to 5:20 pm Artifact Detection Chair: Kathrin Berkner, Ricoh Innovations, Inc. 3:40 pm: Ringing artifact measurement for JPEG images, X. Feng, J. P. Allebach, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-11] 4:00 pm: A hybrid intelligence approach to artifact recognition in digital publishing, F. Vega, H. J. Santos-Villalobos, Univ. de Puerto Rico Mayagьez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-12] 4:20 pm: Nearest-neighbor and bilinear resampling factor estimation to detect blockiness or blurriness of an image, A. Suwendi, J. P. Allebach, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-13] 4:40 pm: Analytical model of skew effect in digital press characterization, M. Qiao, J. P. Allebach, Purdue Univ. . . . . . . . [6076-14] 5:00 pm: Detection and location of very small print defects in real time for high-speed digital printing, G. W. Braudaway, IBM Corp. . . [6076-15] Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details. SESSION 4 Conv. Ctr. Room A7 . . . . . . . . . . . . Tues. 9:30 to 11:20 am Document and Image Presentation Chair: Fabio Vitali, Univ. degli Studi di Bologna (Italy) 9:30 am: New economy, new strategy: digital technology innovations and applications (Invited Paper), N. Raman, Hewlett-Packard Labs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-16] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:10 to 10:40 am 10:40 am: How small should a document thumbnail be?, K. Berkner, Ricoh Innovations, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-17] 11:00 am: Image object adaptation in variable data printing, J. Fan, Hewlett-Packard Labs.; H. Chao, Hewlett-Packard Co. . . . . . . . [6076-18] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . 11:20 am to 1:20 pm
62
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6076 · Conv. Ctr. Room A7 SESSION 5 Conv. Ctr. Room A7 . . . . . . . . . . . . . Tues. 1:20 to 3:00 pm Document Layout Chair: Fernando Vega, Univ. de Puerto Rico Mayagьez 1:20 pm: Laying out the future of final-form digital documents (Invited Paper), D. F. Brailsford, Univ. of Nottingham (United Kingdom) . . [6076-20] 2:00 pm: Intelligent content fitting for digital publishing, X. Lin, HewlettPackard Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-21] 2:20 pm: A total-fit page-breaking algorithm with user-defined adjustment strategies, A. Di Iorio, L. Furini, F. Vitali, Univ. degli Studi di Bologna (Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-22] 2:40 pm: Extensible layout in functional documents, J. Lumley, R. Gimson, O. Rees, Hewlett-Packard Ltd. (United Kingdom) . . . . . [6076-23] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm SESSION 6 Conv. Ctr. Room A7 . . . . . . . . . . . . . Tues. 3:30 to 5:30 pm Publishing, Production, and Workflow Chair: Dhiraj Kacker, Caslon & Co., Inc. 3:30 pm: Production digital printing: making the leap from emerging to everyday (Invited Paper), C. Valiquette, Caslon & Co. . . . . . . . . [6076-24] 4:10 pm: WARP (workflow for automated and rapid publishing): a framework for end-to-end automated digital publishing workflows, P. Joshi, Hewlett-Packard Co. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6076-25] 4:30 pm: A scheduling framework applied to digital publishing workflows, W. Rivera, W. L'Ozano, Univ. de Puerto Rico Mayagьez [607626] 4:50 pm: Desktop binding: a novel approach to booklet making hardware, S. W. Trovinger, Hewlett-Packard Co. . . . . . . . . . . . . . [6076-27] 5:10 pm: Color variance in PDF-based production workflows, M. P. Riordan, Rochester Institute of Technology . . . . . . . . . . . . . . . . . [6076-28] Demonstration Session - Tuesday
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
63
Conference 6077 · Conv. Ctr. Rooms B2 and B3
Tuesday-Thursday 17-19 January 2006 · Proceedings of SPIE Vol. 6077 Visual Communications and Image Processing 2006
Conference Chairs: John G. Apostolopoulos, Hewlett-Packard Labs.; Amir Said, Hewlett-Packard Labs. Program Committee: Kiyoharu Aizawa, The Univ. of Tokyo (Japan); Yucel Altunbasak, Georgia Institute of Technology; Alan C. Bovik, The Univ. of Texas at Austin; Chang Wen Chen, Florida Institute of Technology; Charles D. Creusere, New Mexico State Univ.; Gerard de Haan, Philips Research Labs. (Netherlands); Edward J. Delp III, Purdue Univ.; Eric Dubois, Univ. of Ottawa (Canada); Frederic Dufaux, Emitall S.A. (Switzerland); Touradj Ebrahimi, Emitall S.A. (Switzerland); Onur G. Guleryuz, DoCoMo Communications Labs. USA, Inc.; Sheila S. Hemami, Cornell Univ.; T. Russell Hsing, Telcordia Technologies, Inc.; Lina J. Karam, Arizona State Univ.; Janusz Konrad, Boston Univ.; Alex C. Kot, Nanyang Technological Univ. (Singapore); C.-C. Jay Kuo, Univ. of Southern California; Reginald L. Lagendijk, Technische Univ. Delft (Netherlands); Shipeng Li, Microsoft Research Asia (China); Yi Liang, Qualcomm; Bangalore S. Manjunath, Univ. of California/Santa Barbara; Peyman Milanfar, Univ. of California/Santa Cruz; Sanjit K. Mitra, Univ. of California/Santa Barbara; Antonio Ortega, Univ. of Southern California; Jцrn Ostermann, Univ. Hannover (Germany); Sethuraman Panchanathan, Arizona State Univ.; Thrasyvoulos N. Pappas, Northwestern Univ.; William A. Pearlman, Rensselaer Polytechnic Institute; Fernando M. B. Pereira, Instituto Superior Tйcnico (Portugal); Bйatrice Pesquet-Popescu, Йcole Nationale Supйrieure des Tйlйcommunications (France); Fatih M. Porikli, Mitsubishi Electric Research Labs.; Majid Rabbani, Eastman Kodak Co.; Kannan Ramchandran, Univ. of California/Berkeley; Kenneth Rose, Univ. of California/Santa Barbara; Paul Salama, Indiana Univ./Purdue Univ. at Indianapolis; Dan Schonfeld, Univ. of Illinois at Chicago; Thomas Sikora, Technische Univ. Berlin (Germany); Eckehard G. Steinbach, LudwigMaximilians-Univ. Mьnchen; Robert L. Stevenson, Univ. of Notre Dame; Thomas Stockhammer, Technische Univ. Mьnchen (Germany); Huifang Sun, Mitsubishi Electric Research Labs.; Ming-Ting Sun, Univ. of Washington; Andrew G. Tescher, AGT Associates; Bhaskaran Vasudev, Epson Palo Alto Lab.; Anthony Vetro, Mitsubishi Electric Research Labs.; John W. Woods, Rensselaer Polytechnic Institute Tuesday 17 January Plenary Speaker . . . . . . . . . . . . . . Tues. 8:30 to 9:15 am Marriott Ballroom 1-6 Image Processing: Interconnections Thomas S. Huang, Beckman Institute for Advanced Science and Technology, Univ. of Illinois at Urbana-Champaign See p. 7 for details.
Sessions 1 and 2 run concurrently.
SESSION 1
SESSION 2
Conv. Ctr. Room B3 . . . . . . . . . . . . . Tues. 1:30 to 5:10 pm Special Session: Superresolution Chair: Onur G. Guleryuz, DoCoMo Communications Labs. USA, Inc. 1:30 pm: Superresolution of text from nonideal video (Invited Paper), X. Li, West Virginia Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-01] 1:55 pm: Registration of aliased images for super-resolution imaging (Invited Paper), P. Vandewalle, L. M. Sbaiz, S. E. Sьsstrunk, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland); M. Vetterli, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland) and Univ. of California/ Berkeley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-02] 2:20 pm: A practical approach to superresolution (Invited Paper), S. Farsiu, Univ. of California/Santa Cruz; M. Elad, Technion - Israel Institute of Technology (Israel); P. Milanfar, Univ. of California/Santa Cruz . . . [6077-03] 2:45 pm: Jitter camera: a superresolution video camera (Invited Paper), M. Ben-Ezra, Siemens Corporate Research; A. Zomet, S. K. Nayar, Columbia Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-04] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm 3:40 pm: Face recognition with independent component based superresolution (Invited Paper), O. G. Sezer, Sabanci Univ. (Turkey); Y. Altunbasak, Georgia Institute of Technology; A. Ercil, Sabanci Univ. (Turkey) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-05] 4:05 pm: Toward new a compression standard using superresolution techniques (Invited Paper), R. Molina, Univ. de Granada (Spain); A. K. Katsaggelos, Northwestern Univ.; L. Alvarez, J. Mateos, Univ. de Granada (Spain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-25] 4:30 pm: Robust superresolution based on pixel-level selectivity, Z. A. Ivanovski, L. Panovski, Ss Cyril and Methodius Univ. (Macedonia); L. J. Karam, Arizona State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-07] 4:50 pm: Resolution enhancement of low-quality videos using a highresolution frame, T. Q. Pham, L. J. van Vliet, Technische Univ. Delft (Netherlands); K. Schutte, TNO-FEL (Netherlands) . . . . . . . . . . . [6077-08]
Conv. Ctr. Room B2 . . . . . . . . . . . . . Tues. 1:30 to 4:40 pm Video Coding Chair: Robert L. Stevenson, Univ. of Notre Dame 1:30 pm: Predictive fast motion/disparity search for multiview video coding, P. Lai, A. Ortega, Univ. of Southern California . . . . . . . . . [6077-09] 1:50 pm: Complexity scalable motion estimation for H.264/AVC, C. Kim, Univ. of Southern California; J. Xin, A. Vetro, Mitsubishi Electric Research Labs.; C. C. J. Kuo, Univ. of Southern California . . . . . [6077-10] 2:10 pm: Depth map compression for unstructured lumigraph rendering, U. Fecker, A. Guenegues, I. Scholz, A. Kaup, FriedrichAlexander-Univ. Erlangen-Nьrnberg (Germany) . . . . . . . . . . . . . . [6077-11] 2:30 pm: Shape adaptive integer transform for coding arbitrarily shaped objects in H264/AVC, X. Li, E. A. Edirisinghe, H. E. Bez, Loughborough Univ. (United Kingdom) . . . . . . . . . . . . . . . . . . . . [6077-12] 2:50 pm: Optimal bit allocation for hybrid scalable/multiple-description video transmission over wireless channels, M. Bansal, M. K. Jubran, L. P. Kondi, SUNY/Univ. at Buffalo . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-13] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:10 to 3:40 pm 3:40 pm: Space-time multiple description video coding, D. Wang, N. Canagarajah, D. Bull, Univ. of Bristol (United Kingdom) . . . . . . . . [6077-14] 4:00 pm: Improving sequential decoding of CABAC encoded data via objective adjustment of the complexity-efficiency trade-off, S. BenJamaa, M. Kieffer, P. Duhamel, Univ. Paris XI (France) and Ctr. National de la Recherche Scientifique (France) . . . . . . . . . . . . . . . . . . . . . . . . [6077-15] 4:20 pm: New intra-luma prediction mode in H.264/AVC using collocated weighted chroma pixel value, I. Cho, J. Lee, W. Lee, D. Jeong, Inha Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-16]
64
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6077 · Conv. Ctr. Rooms B2 and B3
Posters and Demonstrations-Tuesday Demonstrations . . . . . . . . . . . . . . . . . . . . . . 5:30 to 8:30 pm A symposium-wide demonstration session will be open to attendees 5:30 to 8:30 pm Tuesday evening in Conv. Ctr. Exhibit Hall 1. Demonstrators will provide interactive, hands-on demonstrations of a wide-range of products related to Electronic Imaging. Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:30 to 7:00 pm Posters will be placed on display after 10:00 am in Exhibit Hall 1. A poster session, with authors present at their posters, will be held Tuesday evening, 5:30 to 7:00 pm. Coupled nonlinear-diffusion color image sharpening based on the chromaticity-brightness model, T. Saito, R. Nosaka, T. Komatsu, Kanagawa Univ. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-56] Parallel implementation of arbitrary-shaped MPEG-4 decoder for multiprocessor systems, M. Pastrnak, P. H. N. de With, LogicaCMG (Netherlands) and Technische Univ. Eindhoven (Netherlands); S. Stuijk, J. van Meerbergen, Technische Univ. Eindhoven (Netherlands) [6077-57] A framework for fast mode decision in the H.264 video coding standard, M. Y. Yang, C. Grecos, Loughborough Univ. (United Kingdom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-58] Interpolation of still images using the decay and persistence properties of discrete wavelet transform coefficients, W. Kwak, R. Park, J. Lee, Sogang Univ. (South Korea) . . . . . . . . . . . . . . . . [6077-59] Optimum computational resource allocation and energy minimization for video encoding on portable devices, Z. He, Univ. of Missouri/Columbia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-60] Optimal video sensing strategy and performance analysis for wireless video sensors, Z. He, Univ. of Missouri/Columbia . . [6077-61] Shot boundary detection using scale invariant feature matching, M. Park, R. Park, S. W. Lee, Sogang Univ. (South Korea) . . . . . . . [6077-62] Extracting focused object from low depth-of-field image sequences, J. Park, C. Kim, Information and Communications Univ. (South Korea) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-63] Adaptive de-blocking filter for low bit rate applications, X. Jin, G. Zhu, Huazhong Univ. of Science and Technology (China) . . . . [6077-64] Fast intra-mode decision algorithm of H.264/AVC, W. Lee, J. Lee, I. Cho, D. Jeong, Inha Univ. (South Korea) . . . . . . . . . . . . . . . . . [6077-65] A splitting algorithm for touched particle based on distance map and particle shape information, W. Wang, Chongqing Univ. of Posts and Telecommunications (China) . . . . . . . . . . . . . . . . . . . . . . . [6077-66] Geo-registration of aerial images by feature matching, Z. Wu, H. Qian, M. Zhu, Zhejiang Univ. (China) . . . . . . . . . . . . . . . . . . . . [6077-67] A new segmentation approach using Gaussian color model and temporal information, M. Karaman, L. Goldmann, T. Sikora, Technische Univ. Berlin (Germany) . . . . . . . . . . . . . . . . . . . . . . [6077-68] Region-based transform-domain video scrambling, F. Dufaux, T. Ebrahimi, Emitall S.A. (Switzerland) and Йcole Polytechnique Fйdйrale de Lausanne (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-69] Robust face detection based on components and their topology, L. Goldmann, U. Mцnich, T. Sikora, Technische Univ. Berlin (Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-70] Resolution scalable SPIHT, D. Choundappan, P. Salama, M. Rizkalla, M. El-Sharkawy, Indiana Univ./Purdue Univ. at Indianapolis . . [6077-72]
Robust global motion estimation in video stabilization for reducing visually induced motion sickness, I. Tsubaki, Kanagawa Univ. (Japan); T. Morita, NHK Engineering Services (Japan); K. Aizawa, The Univ. of Tokyo (Japan); T. Saito, Kanagawa Univ. (Japan) . . . . . . . . . . . [6077-73] Spatially variant morphological image processing: theory and applications, N. Bouaynaya, D. Schonfeld, Univ. of Illinois at Chicago . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-74] Video frame rate up conversion under inconsistent camera, J. Wang, Wayne State Univ.; N. Patel, W. Grosky, Univ. of Michigan/ Dearborn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-75] Pre-compression rate allocation for JPEG2000 encoders in power constrained devices, F. Chebil, R. Kurceren, Nokia Americas [6077-76] Distortion fluctuation control for 3D wavelet-based video coding, V. Seran, L. P. Kondi, SUNY/Univ. at Buffalo . . . . . . . . . . . . . . . . [6077-77] Reduction of MPEG ringing artifacts using adaptive sigma filter, H. Pan, Sharp Labs. of America, Inc. . . . . . . . . . . . . . . . . . . . . . . [6077-78] Bit-stream extraction to maximize perceptual quality using quality information table in SVC, Y. S. Kim, Y. J. Jung, T. C. Thang, Y. M. Ro, Information and Communications Univ. (South Korea) . . . . . . [6077-79] Active surfaces for video tracking and 3D segmentation based on a new method for multidimensional optimization, N. Bouaynaya, D. Schonfeld, Univ. of Illinois at Chicago . . . . . . . . . . . . . . . . . . . [6077-80] Robust transmission of packet-based H.264/AVC video with data partitioning over DS-CDMA wireless channels, A. V. S. Mantravadi, M. Bansal, L. P. Kondi, SUNY/Univ. at Buffalo . . . . . . . . . . . . . [6077-81] A novel VLC-based on second-run-level coding and dynamic truncation, C. h. Cui, W. y. Liu, X. Jin, Huazhong Univ. of Science and Technology (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-82] Efficient coding scheme for super high definition video based on extending H.264 high profile, S. Naito, A. Matsumura, KDDI R&D Labs. (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-83] Fish tracking by combining motion-based segmentation and particle filtering, E. Bichot, L. Mascarilla, P. Courtellemont, Univ. de La Rochelle (France) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-85] Multiresolution color patch extraction, P. Obrador, Hewlett-Packard Labs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-86] Disparity estimation using edge model for stereo video compression, H. J. Kim, Y. Lee, J. B. Ra, K. Cho, Korea Advanced Institute of Science and Technology (South Korea) . . . . . . . . . [6077-87] Adaptive λ estimation in Lagrangian rate-distortion optimization for video coding, L. Chen, Mobilygen Corp. . . [6077-88]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
65
Conference 6077 · Conv. Ctr. Rooms B2 and B3
Wednesday 18 January
Plenary Speaker . . . . . . . . . . . . . . . Wed. 8:30 to 9:15 am Marriott Ballroom 1-6 Computational Imaging Methods for Functional Brain Mapping and Molecular Imaging Richard Leahy, Univ. of Southern California See p. 7 for details.
Sessions 3 and 6 run concurrently.
SESSION 3
SESSION 6
Conv. Ctr. Room B3 . . . . . . . . . Wed. 9:30 am to 12:05 pm Special Session: Current Topics in Video Coding Chairs: Mary L. Comer, Purdue Univ.; Edward J. Delp III, Purdue Univ. 9:30 am: Rate-distortion analysis of SP and SI frames (Invited Paper), E. Setton, P. Ramanathan, B. Girod, Stanford Univ. . . . . . . . . . . . [6077-17] 9:55 am: Wyner-Ziv video coding with universal prediction, Z. Li, L. Liu, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-18] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:20 to 10:50 am 10:50 am: Hybrid scalable video coding with multiple description and layered coding (Invited Paper), G. Zhang, R. L. Stevenson, Univ. of Notre Dame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-19] 11:15 am: Wyner-Ziv video coding: a motion estimation perspective, Z. Li, L. Liu, E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . [6077-20] 11:40 am: A new approach to motion compensation in spatially scalable video coding (Invited Paper), M. L. Comer, Purdue University . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-21] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 1:35 pm
Conv. Ctr. Room B2 . . . . . . . . . Wed. 9:30 am to 12:00 pm Computer Vision Chair: Dan Schonfeld, Univ. of Illinois at Chicago 9:30 am: Parallel multiple target tracking using multiple cooperative trackers, W. Qu, D. Schonfeld, Univ. of Illinois at Chicago; M. A. Mohamed, Motorola, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-31] 9:50 am: Rao-Blackwellised particle filter with adaptive system noise and its evaluation for tracking in surveillance, X. Xu, B. Li, Arizona State Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-32] 10:10 am: Spatial detection of logos as outliers from the content, A. Ekin, Philips Research Labs. (Netherlands) . . . . . . . . . . . . . . . . . [6077-33] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am 11:00 am: Finding corners in images by foveated search, T. L. Arnow, A. C. Bovik, The Univ. of Texas at Austin . . . . . . . . . . . . . . . . . . . . . [6077-34] 11:20 am: Two-dimensional regularized disparity estimation based on the Gabor transform, X. Huang, E. Dubois, Univ. of Ottawa (Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-35] 11:40 am: Plane-based calibration of cameras with zoom variation, C. Yu, G. Sharma, Univ. of Rochester . . . . . . . . . . . . . . . . . . . . . . . . [6077-36] Lunch/Exhibition Break . . . . . . . . . . . . . . . . . . . . . . . . . . 12:00 to 2:00 pm
Sessions 4 and 7 run concurrently.
SESSION 4
SESSION 7
Conv. Ctr. Room B3 . . . . . . . . . . . . . Wed. 1:35 to 3:15 pm Special Session: Advances in Image/Video Coding and Delivery Chair: John G. Apostolopoulos, Hewlett-Packard Labs. 1:35 pm: Distributed sender-driven video streaming (Invited Paper), J. Chakareski, P. Frossard, Йcole Polytechnique Fйdйrale de Lausanne (Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-22] 2:00 pm: Advances in video encoder optimization (Invited Paper, Presentation Only), A. Dumitras, Apple Computer, Inc. . . . . . . . [6077-23] 2:25 pm: Suprathreshold visual psychophysics and structure-based visual masking (Invited Paper), S. S. Hemami, D. M. Chandler, B. Chern, J. A. Moses, Cornell Univ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-24] 2:50 pm: Video compression with flexible playback order based on distributed source coding (Invited Paper), N. Cheung, H. Wang, A. Ortega, Univ. of Southern California . . . . . . . . . . . . . . . . . . . . . . . [6077-89] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:15 to 3:45 pm
Conv. Ctr. Room B2 . . . . . . . . . . . . . Wed. 2:00 to 3:00 pm Video Processing Chair: Remco Muijs, Philips Research Labs. (Netherlands) 2:00 pm: Classification-based hybrid filters for image processing, H. Hu, Technical Univ. of Eindhoven (Netherlands); G. de Haan, Philips Research Labs. (Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-37] 2:20 pm: Solving occlusion in film judder elimination, E. B. Bellers, Philips Semiconductors; J. van Gurp, Philips Semiconductors (Netherlands); J. Janssen, Philips Semiconductors; R. A. C. Braspenning, R. Wittebrood, Philips Research Labs. (Netherlands) . . . . . . . . . . [6077-39] 2:40 pm: Similarity-independent and non-iterative algorithm for subpixel motion estimation, M. Shimizu, S. Chang, M. Okutomi, Tokyo Institute of Technology (Japan) . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-40] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3:00 to 3:30 pm
66
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conference 6077 · Conv. Ctr. Rooms B2 and B3
Sessions 5 and 8 run concurrently.
SESSION 5
SESSION 8
Conv. Ctr. Room B3 . . . . . . . . . . . . . Wed. 3:45 to 5:25 pm Distributed Source Coding Chair: Majid Rabbani, Eastman Kodak Co. 3:45 pm: H.264 redundant slices for systematic lossy error protection of video, S. D. Rane, B. Girod, Stanford Univ. . . . . . . . . . . . . . . . [6077-26] 4:05 pm: Correlation estimation and performance optimization for distributed image compression, Z. He, Univ. of Missouri/ Columbia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-27] 4:25 pm: Correlation structure analysis for distributed video compression over wireless video sensor networks, Z. He, Univ. of Missouri/Columbia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-28] 4:45 pm: Distributed multiview video coding, X. Guo, Harbin Institute of Technology (China) and Microsoft Research Asia (China); Y. Lu, F. Wu, Microsoft Research Asia (China); W. Gao, Institute of Computing Technology (China); S. Li, Microsoft Research Asia (China) and Beijing Institute of Technology (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-29] 5:05 pm: Free viewpoint switching in multiview video streaming using Wyner-Ziv video coding, X. Guo, Harbin Institute of Technology (China); Y. Lu, F. Wu, Microsoft Research Asia (China); W. Gao, Institute of Computing Technology (China); S. Li, Microsoft Research Asia (China) . . . . . [6077-30]
Conv. Ctr. Room B2 . . . . . . . . . . . . . Wed. 3:30 to 4:30 pm Image Coding Chair: William A. Pearlman, Rensselaer Polytechnic Institute 3:30 pm: A wavelet-based two-stage near-lossless coder with Linfinity error scalability, S. Yea, W. A. Pearlman, Rensselaer Polytechnic Institute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-41] 3:50 pm: Region of interest access with three-dimensional SBHP algorithm, Y. Liu, W. A. Pearlman, Rensselaer Polytechnic Institute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-42] 4:10 pm: Optimal JPEG2000 rate control mechanism applicable for super low delay distribution of HDTV programs, S. Naito, A. Koike, Kokusai Dinshin Denwa KK (Japan) . . . . . . . . . . . . . . . . . . . . . . . [6077-43]
Thursday 19 January SESSION 9 Conv. Ctr. Room B3 . . . . . . . . . . . Thurs. 9:30 to 11:40 am Media over Networks Chair: Paul Salama, Indiana Univ./Purdue Univ. at Indianapolis 9:30 am: Error resilience in network driven Wyner-Ziv video coding, L. Liu, P. Sabria, Purdue Univ.; J. Prades-Nebot, Univ. Politиcnica de Valиncia (Spain); E. J. Delp III, Purdue Univ. . . . . . . . . . . . . . . . . . . . . . . . . [6077-45] 9:50 am: Receiver buffer requirement for video streaming over TCP, T. Kim, Freescale Semiconductor, Inc.; M. H. Ammar, Georgia Institute of Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-46] 10:10 am: Effective overlay multicast tree constructing algorithm over multiple differentiated-service networks, D. B. Lee, H. Song, Pohang Univ. of Science and Technology (South Korea) . . . . . . . . . . . . . . [6077-47] Coffee Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10:30 to 11:00 am 11:00 am: A novel source rate control algorithm for video streaming over the Internet, Z. Peng, Tsinghua Univ. (China); W. Zeng, Univ. of Missouri/Columbia; C. W. Li, Tsinghua Univ. (China) . . . . . . . . . . [6077-48] 11:20 am: Motion-embedded residual error for packet loss recovery of video transmission and encryption, S. Sun, Institute of Information Science (Taiwan); J. Chen, National Central Univ. (Taiwan); C. Lu, Institute of Information Science (Taiwan); P. Chang, K. Fan, National Central Univ. (Taiwan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-49] Lunch Break . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:40 am to 1:30 pm
SESSION 10 Conv. Ctr. Room B3 . . . . . . . . . . . . Thurs. 1:30 to 3:10 pm Scalable Video Coding Chair: John W. Woods, Rensselaer Polytechnic Institute 1:30 pm: A new structure of 3D dual-tree discrete wavelet transforms and applications to video denoising and coding, F. Shi, B. Wang, I. W. Selesnick, Y. Wang, Polytechnic Univ. . . . . . . . . . . . . . . . . . . . . . [6077-50] 1:50 pm: On scalable lossless video coding based on subpixel accurate MCTF, S. Yea, W. A. Pearlman, Rensselaer Polytechnic Institute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-51] 2:10 pm: Aliasing reduction via frequency roll-off for scalable image/ video coding, Y. Wu, J. W. Woods, Rensselaer Polytechnic Institute . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-52] 2:30 pm: Adaptive in-band motion compensated temporal filtering based on motion mismatch detection in the high-pass subbands, A. Gao, N. Canagarajah, D. Bull, Univ. of Bristol (United Kingdom) . [6077-53] 2:50 pm: Quality-fluctuation-constrained rate allocation for MCTFbased video coding, Y. Chen, Shanghai Jiao Tong Univ. (China); J. Xu, F. Wu, Microsoft Research Asia (China); H. Xiong, Shanghai Jiao Tong Univ. (China) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . [6077-54]
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
67
Technical Abstract Summaries
6055A Stereoscopic Displays and Applications XVII (Woods, Dodgson, Merritt) . . . . . . . . . . . . . . . . . . p. 69 6055BThe Engineering Reality of Virtual Reality 2006 (Bolas, McDowall) . . . . . . . . . . . . . . . . . . . . . . . . . p. 82 6056 Three-Dimensional Image Capture and Applications VI (Corner, Li, Tocheri) . . . . . . . . . . p. 86 6057 Human Vision and Electronic Imaging XI (Rogowitz, Pappas, Daly) . . . . . . . . . . . . . . . . . . . p. 95 6058 Color Imaging XI: Processing, Hardcopy, and Applications (Eschbach, Marcu) . . . . . . . . . . . . p. 105 6059 Image Quality and System Performance III (Cui, Miyake) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p.113 6060 Visualization and Data Analysis 2006 (Erbacher, Roberts, Grцhn, Bцrner) . . . . . . . . . . . . . . . . . . . p. 122 6061 Internet Imaging VII (Gevers, Santini, Schettini) p. 129 6062 Spectral Imaging: Eighth International Symposium on Multispectral Color Science (Rosen, Imai, Tominaga) . . . . . . . . . . . . . . . . . . . p. 136 6063 Real-Time Image Processing III (Kehtarnavaz, Laplante) . . . . . . . . . . . . . . . . . . . p. 142 6064A Image Processing: Algorithms and Systems V (Dougherty, Astola, Egiazarian) . . . . . . . . . . . . . . . p.148 6064BApplications of Neural Networks and Machine Learning in Image Processing X (Nasrabadi, Rizvi) . . . . . . . . . . . . . . . . . . . . . . . . p. 158 6065 Computational Imaging IV (Bouman, Miller, Pollak) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 162 6066 Vision Geometry XIV (Latecki, Mount, Wu) . . . . p. 172 6067 Document Recognition and Retrieval XIII (Taghva, Lin) . . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 178 6068 Sensors, Cameras, and Systems for Scientific/Industrial Applications VIII (Blouke) p. 183 6069 Digital Photography II (Sampat, DiCarlo, Martin) p. 189 6070 Machine Vision Applications in Industrial Inspection XIV (Meriaudeau, Niel) . . . . . . . . . . . p. 194 6071 Multimedia Computing and Networking 2006 (Chandra, Griwodz) . . . . . . . . . . . . . . . . . . . . . . p. 201 6072 Security, Steganography, and Watermarking of Multimedia Contents VIII (Delp, Wong) . . . . . . p. 206 6073 Multimedia Content Analysis, Management, and Retrieval 2006 (Chang, Hanjalic, Sebe) . . . . . . . p. 221 6074 Multimedia on Mobile Devices II (Creutzburg, Takala, Chen) . . . . . . . . . . . . . . . . . . . . . . . . . . . p. 229 6075 Optical Security and Counterfeit Deterrence Techniques VI (van Renesse) . . . . . . . . . . . . . . p. 238 6076 Digital Publishing (Allebach, Chao) . . . . . . . . . p. 248 6077 Visual Communications and Image Processing 2006 (Apostolopoulos, Said) . . . . . p. 253
68
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055A: Stereoscopic Displays and Applications XVII
Monday-Wednesday 16-18 January 2006 Part of Proceedings of SPIE Vol. 6055 Stereoscopic Displays and Virtual Reality Systems XIII
6055A-01, Session 1 The use of stereoscopic visualization in chemistry and structural biology M. Husбk, Institute of Chemical Technology Prague (Czech Republic) The results of experimental measurements of molecular structures (such as X-ray diffractions techniques or NMR techniques) are often hard to interpret. The obtained 3D data as electron density maps, atoms coordinates and molecular space arrangements are complex objects and they require sophisticated visualization methods. In the presentation will be given an overview of existing chemical software supporting stereoscopic visualization. The main parts of the lecture will disuse our experiences with creating an ergonomically comfortable working environment for everyday stereoscopic visualization use for chemical structure analysis purpose. Implementation of the necessary stereoscopic visualization functionalities in MCE code developed by us (code for interpretation of X-ray diffraction and quantum mechanical calculations) will be discussed as well. 6055A-02, Session 1 Using stereoscopic real-time graphics to shorten training time for complex mechanical tasks F. Tecchia, M. Carrozzino, F. Rossi, M. Bergamasco, Scuola Superiore Sant'Anna (Italy); M. Vescovi, SIG Simonazzi (Italy) The present paper presents the hands-on results of the use of a large screen stereoscopic installation to train technicians on maintenance tasks of large machineries for a leading mechanical industry. Such machinery, deployed from the company in remote locations around the world, need complex and lengthy maintenance procedures, to be performed periodically by teams of highly trained technicians. The firm organise continuous training classes to a large number of its technicians, using qualified trainers and continuously updating machinery documentation, resulting in long and expensive periods of time of technicians inactivity. Classes involve training on assembly and disassembly operations of the company complex mechanical products and were traditionally based on the use of video documentation, 2D mechanical drawings and live demonstrations on real equipment. In an attempt to improve this process, the firm equipped one of their training centres with a large stereoscopic projection facility and dedicated software, introducing the use of real-time stereo rendering of CAD models for virtual disassembly/assembly sequences. The firm investigated then potential benefits of the new methodology compared to the traditional one. The present article presents an overview of the technological framework used, and summarises the results of such comparison performed over a period of 6 months. 6055A-03, Session 1 Stereoscopic display of 3D models for design visualization K. J. Gilson, Parsons Brinckerhoff Advances in display technology and 3D design visualization applications have made real-time stereoscopic visualization of architectural and engineering projects a reality. Parsons Brinkerhoff (PB) is a transportation consulting firm that has used digital visualization tools from their inception and has helped pioneer the
application of those tools to large scale infrastructure projects. PB is one of the first Architecture/Engineering/Construction (AEC) firms to implement a CAVE- an immersive presentation environment that includes stereoscopic rear-projection capability. The firm also employs a portable stereoscopic front-projection system, and shutter-glass systems for smaller groups. PB is using commercial real-time 3D applications in combination with traditional 3D modeling programs to visualize and present large AEC projects to planners, clients and decision makers in stereo. These presentations create more immersive and spatially realistic presentations of the proposed designs. This paper will present the basic display tools and applications, and the 3D modeling techniques PB is using to produce interactive stereoscopic content. The paper will discuss several architectural and engineering design visualizations we have produced. 6055A-04, Session 1 Stereoscopic image production: live, CGI, and integration E. Criado, Enxebre Entertainment (Spain) Stereoscopic image production is complex, but even more than shooting technical parameters, some of them also considered here, I analyze some other differences between classic or flat production compared to stereoscopic production, that should be taken in care and could make the difference between success and ruin, for a 3-D film. This reading does not try to be an exhaustive analysis, but it could offer a global approach to the problematic of a stereoscopic production, reviewing preproduction , the shooting and postproduction processes, including some details about stereoscopic live and computer graphics imagery integration. 6055A-05, Session 1 Cosmic cookery: making a stereoscopic 3D animated movie N. S. Holliman, C. Baugh, C. Frenk, A. Jenkins, B. Froner, D. Hassaine, J. Helly, N. Metcalfe, T. Okamoto, Univ. of Durham (United Kingdom) This paper describes our experience making a short stereoscopic movie visualizing the development of structure in the universe during the 13.7 billion years from the Big Bang to the present day. Aimed at a general audience for the Royal Society's 2005 Summer Science Exhibition, the movie illustrates how the latest cosmological theories based on dark matter and dark energy are capable of producing structures as complex as spiral galaxies and allows the viewer to compare observations from the real universe with theoretical results. 3D is an inherent feature of the cosmology data sets and stereoscopic visualization provides a natural way to present the images, in addition allowing researchers to visualize these complex data sets. The presentation of the movie used passive polarized projection onto a 2m wide screen and also played back on a Sharp RD3D and subsequently in anaglyph projection at venues without stereoscopic displays. We conclude that the generation of high quality stereoscopic movie content using desktop tools and equipment is feasible. This does require careful quality control and manual intervention, but we believe these overheads are worthwhile when presenting inherently 3D data, as the result is significantly increased impact and better understanding of complex 3D scenes.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
69
Conf. 6055A: Stereoscopic Displays and Applications XVII
6055A-06, Session 2 Evaluation of stereoscopic medical video content on an autostereoscopic display for undergraduate medical education J. F. R. Ilgner, Univ. Hospital Aachen (Germany); T. Kawai, T. Shibata, T. Yamazoe, Waseda Univ. (Japan); M. Westhofen, Univ. Hospital Aachen (Germany) Introduction: This study evaluates the usefulness of stereoscopic video in teaching undergraduate medical students. Material and methods: We chose two clips each of three different microsurgical operations. This material was added by 23 clips of a cochlear implantation, which was specifically edited for a portable computer with an autostereoscopic display (SHARP 3DRD). The footage was edited stereoscopically at the Waseda University by means of our original software for non-linear editing of stereoscopic 3-D movies. 25 4th year medical students who participated in the general ENT course at Aachen University Hospital were asked to estimate depth clues within the six video clips plus cochlear implantation clips. Another 25 4th year students who were shown the material monoscopically on a conventional laptop served as control. Results: The monoscopic group generally estimated resection depth to much lesser values than in reality. Although this was the case with some participants in the stereoscopic group, too, the estimation of depth features reflected the enhanced depth impression provided by stereoscopy. Conclusion: Following first implementation of stereoscopic video teaching, medical students who are inexperienced with ENT surgical procedures are able to reproduce depth information and therefore anatomically complex structures to a greater extent following stereoscopic video teaching. 6055A-07, Session 2 Stereoscopic visualization and editing of automatic abdominal aortic aneurysms (AAA) measurements for stent graft planning L. Zhou, Y. P. Wang, C. Goh, R. Kockro, L. Serra, Volume Interactions Pte. Ltd. (Singapore) For stent graft selection in the treatment of abdominal aortic aneurysms (AAA) anatomic considerations are important. They determine GO/NO-GO of the treatment and help customize the stent. Current systems for AAA stent insertion planning based on pre-operative CT and MR of the patient do not provide an intuitive interface to view the resulting measurements against the preoperative CT/MR. Subsequent modifications of the measurements are frequent when automatic algorithms are inaccurate. However, 3D editing is difficult to achieve because of the limitations of monoscopic displays and 2D interface. In this paper, we present a system for automatic AAA measurement and interactive 3D editing. The strength of this approach is that the resulting measurements can be reviewed and edited interactively in the 3D context of the volumetric rendering of the aorta, so that relationships of the measurements and the aorta are clearly perceived. This understanding is facilitated by the stereoscopic rendering that makes it possible to see the transparent vessel and its corresponding measurements all in one image.
6055A-08, Session 2 A hybrid virtual environment for training of radiotherapy treatment of cancer. R. Phillips, The Univ. of Hull (United Kingdom) There is often insufficient access to patients and linear accelerator treatment rooms to train radiotherapy students. An alternative approach is to train using a hybrid Immersive Visualization Environment (IVE) that simulates an actual radiotherapy treatment machine controlled with the actual machine handheld control pendant. A study of training using this IVE is presented for "skin apposition" treatment, where the patient couch and radiotherapy equipment are positioned so that the X-ray beam strikes the skin perpendicularly. The IVE developed comprises a virtual treatment room with a linear accelerator, modelled from laser scan data, stereoscopically projected onto a 16 x 8 foot work-wall. A genuine linear accelerator control handheld "pendant" provided the user interface to the virtual linear accelerator thus creating a hybrid IVE. A virtual patient, based on the visible human female dataset, complete with rectangular markings for a range of different treatment sites, provided a range of treatment scenarios. Quantitative metrics on how well the patient was positioned were also produced by the IVE. A study of 42 students was conducted to evaluate learning. 93% of students perceived an improvement in their understanding of this treatment using the IVE and 69% found the control system to be "easy to master". 6055A-09, Session 2 Blur spot limitations in distal endoscope sensors A. Yaron, Visionsense Inc.; M. Shechterman, N. Horesh, U. Ronen, Visionsense Ltd. (Israel) With current development of sensors, displays and bandwidth, image quality of video systems is typically limited by optics. Further reduction of pixel size in distal endoscope CCD or CMOS cameras will not provide a better image quality. The limiting factor is the blur spot (diffraction limit) phenomena. Stereo sensors that multiplex the horizontal field are indifferent to this optical effect and thus can provide additional depth resolution without sacrificing 2D resolution. Additional advantages of such sensors are reduced aliasing and improved tolerance to noise. 6055A-10, Session 3 Visual comfort with mobile stereoscopic gaming J. P. Hдkkinen, Nokia Research Ctr. (Finland) and Univ. of Helsinki (Finland); M. Liinasuo, Nokia Research Ctr. (Finland); J. Takatalo, G. S. Nyman, Univ. of Helsinki (Finland) Stereoscopic displays are known to cause eye strain because of the convergence-accommodation conflict caused by the stereoscopic content. We tested the level of eye strain with a stereoscopic mobile phone game with small disparities. The users played a simple puzzle game with Sharp SH505i mobile phone which has a parallax barrier autostereoscopic display. There were 5 experiments with different disparity in each: 1) Far depth 5.4 arc min, 2) Far depth 2.7 arc min, 3) Zero disparity, 4) Near depth 2.7 arc min and 5) Near depth 5.4 arc min. In every experiment there were 30 users, so the total number of participants was 150. We measured subjective sickness symptoms with the Simulator Sickness Questionnaire and visual functioning by testing the heterophoria and the near point of accommodation. The results showed that most of the users got
70
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055A: Stereoscopic Displays and Applications XVII
some eye strain symptoms both in the disparity conditions and in the zero disparity condition. The mean eye strain levels did not significantly differ in the zero disparity and other experimental conditions. The results indicate that using small disparities in the mobile stereoscopic displays enables a comfortable user experience. 6055A-11, Session 3 Effect of disparity and motion on visual comfort of stereoscopic images F. Speranza, J. W. Tam, R. Renaud, Communications Research Ctr. Canada (Canada); N. Hur, Electronics and Telecommunications Research Institute (South Korea) It is well known that some viewers experience visual discomfort when looking at stereoscopic displays. One of the factors that can give rise to visual discomfort is the presence of large horizontal disparities. The relationship between excessive horizontal disparity and visual comfort has been well documented for the case in which disparity magnitude does not change across space and time, e.g. for objects in still images. Much less is known about the case in which disparity magnitude varies over time, e.g., objects moving in depth at some velocity. In this study, we investigated the relationship between binocular disparity, object motion and visual comfort using computer-generated stereoscopic video sequences. Specifically, viewers were asked to rate the visual comfort of stereoscopic sequences that had objects moving periodically back and forth in depth. These sequences varied with respect to the number, size, position in depth, and velocity of movement of the objects included in the scene. The results indicated that with motion in depth, ratings of visual comfort depend on the velocity of the stereoscopic objects rather than on disparity per se. The results also suggest that rapid switches between crossed and uncrossed disparities might negatively affect visual comfort. 6055A-12, Session 3 Analysis of an autostereoscopic display: the perceptual range of the threedimensional visual fields and saliency of static depth cues P. R. Havig, Air Force Research Lab.; J. P. McIntire, Consortium Research Fellows Program; R. McGruder, U.S. Air Force Academy Autostereoscopic displays offer users the unique ability to view 3dimensional (3D) imagery without special eyewear or headgear. However, the users' head must be within limited "eye boxes" or "viewing zones." Further, little research has evaluated these viewing zones from a human-in-the-loop, subjective perspective. In the first study, twelve participants evaluated the quality and amount of perceived 3D images. We manipulated distance from observer, viewing angle, and stimuli to characterize the perceptual viewing zones. The data was correlated with objective measures to investigate the amount of concurrence between the objective and subjective measures. In a second study we investigated the benefit of generating stimuli that take advantage of monocular depth cues. The purpose of this study was to determine if one could develop optimal stimuli that would give rise to the greatest 3D effect with offaxis viewing angles. Twelve participants evaluated the quality of depth perception of various stimuli each made up of one monocular depth cue (i.e., linear perspective, occlusion, haze, size, texture, and horizon). Viewing zone analysis is discussed in terms of optimal viewing distances and viewing angles. Stimuli properties are discussed in terms of image complexity and depth cues present.
6055A-13, Session 3 Effects of gender, application, experience, and constraints on interaction performance using autostereoscopic displays Z. Y. Alpaslan, S. Yeh, A. A. Rizzo III, A. A. Sawchuk, Univ. of Southern California We describe a set of experiments that compare 2D CRT, shutter glasses and autostereo displays; measure user preference for different tasks in different displays; measure the effect of previous user experience in the interaction performance for new tasks; and measure the effect of constraining user's hand motion and hand eye coordination. In this set of tests, we used interactive object selection and manipulation tasks using standard scalable configurations of 3D block objects. We also used a 3D depth matching test in which subjects are instructed to align two objects located next to each other on the display to the same depth plane. New subjects tested with hands out of field of view constraint performed more efficiently with glasses than with autostereoscopic displays, meaning they were able to match the objects with less movement. This constraint affected females more negatively than males. From the results of the depth test, we note that previous subjects on average performed better than the new subjects. Previous subjects had more correct results than the new subjects, and they finished the test faster than the new subjects. The depth test showed that glasses are preferred to autostereo displays in a task that involves only stereoscopic depth. 6055A-14, Session 3 Examination of asthenopia recovery using stereoscopic 3D display with dynamic optical correction T. Shibata, T. Kawai, K. Ohta, L. Jae Lin, Waseda Univ. (Japan); M. Otsuki, N. Miyake, Nikon Corp. (Japan); Y. Yoshihara, Arisawa Manufacturing Co., Ltd. (Japan); T. Iwasaki, Univ. of Occupational and Environmental Health (Japan) The common cause of asthenopia is viewing objects from a short distance, such as in VDT (Visual Display Terminal) work. In general, recovery from asthenopia, especially accommodative asthenopia, is aided by looking into the distance. The authors have developed a stereoscopic 3-D display with dynamic optical correction. The display can reduce the discrepancy between accommodation and convergence while viewing stereoscopic 3-D images. Since the display presents images as if they were actually in the distance, the authors considered that viewing images on the display might reduce asthenopia. The authors previously performed the visual acuity test before and after presenting stereoscopic 3-D images in order to verify the hypothesis. The result showed that there was a tendency to recover asthenopia. In this study, the authors developed a feedback function of the presenting stereoscopic 3-D images on the developed display for the purpose of making viewer's distant vision more effective. Using the function, the refraction under viewing is fed back for the image presentation, and the viewer can gradually take distant vision. From the result of the experiment, it was suggested that asthenopia recovery using stereoscopic 3-D images was effective in comparison with 2-D images.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
71
Conf. 6055A: Stereoscopic Displays and Applications XVII
6055A-15, Session 4
6055A-18, Session 4
High-resolution insets in projector-based stereoscopic displays: principles and techniques G. Godin, P. Massicotte, L. Borgeat, National Research Council Canada (Canada) We propose a dual-resolution foveated stereoscopic display built from commodity projectors and computers. The technique is aimed at increasing the access to fine details of 3D models on a wall-type interactive display: it projects a high-resolution inset (or fovea, by analogy to biological vision) that is registered in image space with the large display. A specific issue that must be addressed is the conflict in depth perception between the apparent depth of the natural boundary of the projected inset images (visible due to changes in color, brightness, and resolution) and that of the underlying scene being displayed. We solve this problem by dynamically adjusting the position of the boundaries so that they lie on identical scene points or points visible in only one image. The system can accommodate approximately aligned projectors, through image warping applied as part of the rendering pipeline. The method for image boundary adjustment is discussed along with implementation details. Finally, we show applications of the technique that we are exploring as part of our research in collaborative visualization of large 3D models of environments and sites, built using laser range sensing and/or photogrammetry. 6055A-16, Session 4 Stereo projection using interference filters H. Jorke, M. Fritz, Infitec GmbH (Germany) Conventional stereo display techniques by projection separate images for the left and right eye by (i) wavelength multiplexing (this is know also as the classic anaglyph approach, using for instance red-green filters), by (ii) polarization (linear or circular polarization) or by (iii) time multiplexing (shutter glasses technique). Stereo projection using interference filters is an advanced wavelength multiplexing approach, that specifically takes into account the nature of the human eye, which is characterized by three types of receptors, which are associated to the primary colours blue, green and red. Correspondingly, the two filters used for the left and for the right eye have three narrow transmission bands, respectively . The three transmission bands B1, G1 and R1 of the filter type A for the left eye image and the three transmission bands B2, G2 and R2 of the filter type B for the right eye image are placed such in the visible spectral range (400 to 700 nm) that (1) conjugated transmissions bands (B1-B2, G1-G2, R1-R2) are within the sensitivity range of the respective receptor and (2) conjugated transmission bands do not overlap. Advantages of stereo projection using the interference filter technique (Infitec) are: - full colour capacity - superior channel separation - passiveness of glasses - compatibility with any white screen for mobile and portable presentations - compatibility with standard cinema screens - compatibility also with low gain screens (rear and front projection) for superior image homogeneity, especially in tiled display systems. - compatibility with any digital projection technology (LCD, DLP and D-ILA).
3D in digital cinema W. J. Husak, Dolby Labs. The transition from celluloid based cinema to Digital Cinema offers opportunities to the film maker and the cinema operator to provide 3D content to theatrical audiences. In the past, 3D content required anaglyptic glasses or multiple projectors for stereoscopic presentation. Digital Cinema projectors offer the capability to present both monoscopic and stereoscopic content with a minimal amount of infrastructure upgrade. 3D films have been released periodically throughout the history of filmmaking. The releases have ranged from classic horror movies to IMAX documentaries. 3D movies have been considered a niche market with audience acceptance waxing and waning throughout the years. Much of the difficulties have arisen due to limitations in technology and infrastructure. Early film based 3D presentation was using anaglyph. The resulting image was difficult to watch for the length of a standard movie. A recently released film used a method of a mixing of anaglyph and monoscope between scenes. Another method of theatrical 3D is display of two simultaneous orthogonally polarized images using two separate projectors. The dual projectors are expensive and require complementary left and right film reel sets. Digital Cinema projectors have the capability of changing polarization on every other frame allowing the use of a single projector. This capability coupled with appropriately polarized glasses allows full color stereoscopic images with significantly less artifacts than traditional anaglyph presentation. In addition, modern image processing allows conversion of films originally shot in 2D to be converted into 3D. This will allow movies that were never intended for stereoscopic display to be re-released in 3D. Even with digital technologies, there are several issues that need to be overcome in order for 3D cinema to be successfully deployed. Among these issues are viewer fatigue, narrow viewing angles, limited head tilt, and light output. There are also practical issues such as acquisition, distribution, and collection of glasses. This paper will review the opportunities, technologies, and issues related to 3D distribution and presentation using a Digital Cinema infrastructure. 6055A-55, Session 4 Development of the real-time stereoscopic error corrector and convergence controller S. Nam, C. Park, Korean Broadcasting System (South Korea); Y. S. Yu, K. Lee, TVLogic Co. Ltd. (South Korea) The geometric differences between left and right images are known as a main factor of eye fatigue in the stereoscopic system, so the discrepancy should be eliminated for good stereoscopic images. We developed a real-time stereoscopic error corrector which can adjust the vertical errors, the disparity, and the size (field of view) errors of HD moving pictures in VCR tape. The main idea of this system is to extract the common areas in both images by cropping the left and right images independently. For this system, we developed real-time HD scaling hardware and stereoscopic error correcting software. We tested the system with the video streams taken by our HD stereoscopic camera. As a result, we confirmed that the developed system could reduce the efforts and time for correcting the stereoscopic errors compared to the other methods. In this paper, we introduced the mechanism and function of the realtime stereoscopic error corrector developed by KBS, and compared it with non-real-time NLE-based compensation system. Also we evaluated the possibility of using the same hardware as a real-time convergence controller of parallel-axis stereoscopic camera. It doesn't need the parallax control motors which are used to slide the lenses apart from the cameras in moving parallel axis type. We also described the results of it in this paper.
72
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055A: Stereoscopic Displays and Applications XVII
6055A-19, Session 5 Platelet-based coding of depth maps for the transmission of multiview images Y. Morvan, D. Farin, P. H. N. de With, Technische Univ. Eindhoven (Netherlands) Emerging 3-D displays show simultaneously several views of the scene. A direct transmission of some of these views is impractical since various displays support a different number of views and the decoder has to interpolate the intermediate views. The interpolation of views can be simplified by only transmitting the texture data for the central view and corresponding depth map. Additionally to the coding of the texture data, this technique requires the efficient coding of depth maps. Since the depth map represents the scene geometry, sharp edges, corresponding to object boundaries, should be preserved. We propose an algorithm that models depth maps using piecewise linear functions (platelets). To adapt to varying scene detail, we employ a quadtree decomposition that divides the image into rectangular sub-images of variable size, each sub-image being approximated by one platelet. In order to preserve sharp object boundaries, the support area of each platelet is adapted to the object boundary. The subdivision of the quadtree and the selection of the platelet type are optimized in a unified rate-distortion framework. EXPERIMENTAL RESULTS show that the described method can improve the compression of depth maps by 2 dB when compared to a JPEG2000 encoder. 6055A-20, Session 5 Efficient view synthesis from uncalibrated stereo R. A. C. Braspenning, M. Op de Beeck, Philips Research Labs. (Netherlands) 3D Displays that are capable of presenting an illusion of a real 3dimensional scene to the viewer have been a topic of research for a long time. Common approaches include displays that require the viewer to wear glasses in order to separate the left and right image from the mixed light output of the display. Well-known are the redcyan or polarized glasses used for viewing 3D movies in amusement parks. However, wearing such, non-customized, glasses can cause discomfort, especially for longer viewing periods, as would be the case for 3D TV applications. Therefore, so-called auto-stereoscopic displays that do not require the viewer to wear glasses, have been researched. The two main techniques use, either barriers [1], or lenses (lenticular sheets) [2] to create 2 or more views. In this paper, we will focus on the multiview 3D TV application. Using more than 2 views enables a considerable freedom of movement by the viewer and the viewing by multiple viewers simultaneously. Both are very important requirements for the 3D TV application. However, besides the 3D displays, also 3D content needs to be available. Especially the multiview 3D displays still need algorithms to convert existing video material into multiview material delivering a 3D experience. The available video content can be roughly divided into two categories, regular 2D video content and stereo video material, although some multiview content is appearing now. The conversion from 2D material to multiview usually comprises estimating a depth map per pixel first [3] and subsequently rendering new views using the original image and the depth map [4]. In this paper we will only focus on stereo material and use both left and right images to generate new views, without calculating an explicit depth map.
6055A-21, Session 5 A fast image multiplexing method robust to viewer's position and lens misalignment in lenticular 3D displays Y. Lee, J. B. Ra, Korea Advanced Institute of Science and Technology (South Korea) Among various autostereoscopic display systems, the lenticular display is one of the most popular systems due to its easy manufacturability. For N-view lenticular display, N view images are to be regularly sub-sampled and interleaved to produce a 3D image. In our previous work, we have pointed out the two problems that cause 3D image distortion. Namely, the system provides the best quality only at a fixed optimal viewing distance and the lenticular sheet may not be precisely aligned on the LCD pixel array. Then, we have proposed the compensation algorithm to solve these problems. However, the proposed algorithm requires a considerable computational burden. In this paper, we propose a new fast multiplexing algorithm using a mapping table. The mapping table consists of two terms related with a viewer's position and the alignment error, respectively. Since the latter term is fixed for a given display system, we can reduce the computational complexity by repeatedly reusing it. In contrary to the previous algorithm, the proposed algorithm can make real time compensation possible without degrading image quality. 6055A-22, Session 6 Real-time rendering for multiview displays R. M. Berretty, F. J. Peters, Philips Research Labs. (Netherlands); G. Volleberg, Philips Applied Technologies (Netherlands) In video systems, the introduction of 3D video might be the next revolution after the introduction of color. Nowadays multiview autostereoscopic displays are in development. The various views produced on the display differ with respect to their associated camera positions. A video format that is highly suited for rendering from different camera positions is the usual 2D format enriched with a depth related channel, i.e., for each pixel in the video not only its color is given, but also, e.g., its distance to a camera. This format, also called the 2.5D video format, has several advantages. It it is device indepent --it can be used for any multiview system without specifying the number of views at recording time-- and it allows for efficient coding. In this paper, we demonstrate that high image quality can be achieved when using the image plus depth format for 3D video systems. We present an efficient high quality real time rendering algorithm that uses forward mapping to reduce aliasing artefacts. It deals properly with occlusions. Image quality is achieved by advanced filtering techniques, taking into account the optical characteristics of the display. 6055A-23, Session 6 Anisotropic scene geometry resampling with occlusion filling for 3DTV applications J. Kim, T. Sikora, Technische Univ. Berlin (Germany) Image-based rendering and video-based rendering technologies have recently attracted attention as key technologies to provide a dynamic photorealistic environment with depth perception. Coarse samples of the scene geometry cause aliasing problems in uncovered areas with hidden information. To fill the uncovered area, the hidden information should be additionally analyzed to refine the samples for a better rendering quality. In this paper, we investigate a resampling method based on diffusion tensor theory to raise the density of samples. A scene can be divided into several
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
73
Conf. 6055A: Stereoscopic Displays and Applications XVII
homogeneous regions by edges of depth or disparity. However, inaccurate depth or disparity should be localized by edges of texture regions. Resampling condition is defined within causality of edges throughout the evolution in coarse-to-fine scheme using Gaussian scale-space. The variations in texture and scene geometry are incorporated to find real edges of discontinuity. Anisotropic sampling is proceeded using different weighting to suppress the amount of diffusion for orthogonal direction of the real edges, while isotropic sampling is proceeded in the scale-space. Our method propagates the samples over the occlusion before the real edge but avoids local minima of diffusion on large texture gradients inside the edge. Simulation results show improvements in coding efficiency and rendering quality. 6055A-24, Session 6 Distributed rendering for multiview parallax displays T. Annen, Max-Planck-Institut fьr Informatik (Germany); W. Matusik, H. Pfister, Mitsubishi Electric Research Labs.; H. Seidel, Max-Planck-Institut fьr Informatik (Germany); M. Zwicker, Massachusetts Institute of Technology 3D display technology holds great promise for the future of television, virtual reality, entertainment, and visualization. Multiview parallax displays deliver stereoscopic views without glasses to arbitrary positions within the viewing zone. These systems must include a high-performance and scalable 3D rendering subsystem in order to generate multiple views at real-time frame rates. This paper describes a distributed rendering system for large-scale multiview parallax displays built with a network of PCs, commodity graphics accelerators, multiple projectors, and multiview screens. The main challenge is to render various perspective views of the scene and assign rendering tasks effectively. In this paper we investigate two different approaches: Optical multiplexing for lenticular screens and software multiplexing for parallax barrier displays. We describe the construction of large-scale multi-projector 3D display systems using lenticular and parallax-barrier technology. We have developed different distributed rendering algorithms using the Chromium stream-processing framework and evaluate the trade-offs and performance bottlenecks. Our results show that Chromium is well suited for interactive rendering on multiview parallax displays. 6055A-25, Session 7 On the number of viewing zones required for head-tracked autostereoscopic display N. A. Dodgson, Univ. of Cambridge (United Kingdom) A head-tracked display could be made from a two-view autostereoscopic display where head tracking allows the display to swap the two views when the eyes move from viewing zone to viewing zone. Variations in human interpupillary distance mean that this basic two-view version will not work well for the significant minority of the population who have eye separation significantly different from the average. Woodgate et al. proposed, in 1997, that a three-view system would work well. Analysis of an ideal version of their proposal shows that it does work well for the vast majority of the population. However, most multi-view, multi-lobe autostereoscopic displays have drawbacks which mean that, in practice, such a system would be unacceptable because of the inter-view dark zones generated by the inter-pixel dark zones on the underlying display technology. Variations of such displays have been developed which remove the inter-view dark zones by allowing adjacent views to overlap with one another: the views appear to smoothly blend from one to the next at the expense of a little blurring. Such displays need at least five viewing zones to accommodate the majority of the adult population with headtracking and at least six viewing zones to accommodate everyone.
6055A-26, Session 7 Multiview LCD wall system I. Relke, Opticality GmbH (Germany) The principle construction of the unique autostereoscopic 3D LCD wall is considered. This glasses-free 3D LCD wall provides presentation of high-quality stereo images for many users simultaneously. The technical characteristics of the 3D LCD wall are compared with the corresponding parameters of the multiview 3D projection wall. The general equation for the evaluation of the multiview stereo image in this 3D LCD wall will be presented and all of its parameters are analysed. We introduce here the fundamental matrices, in which will be contained the information about the contribution of the different views in every subpixel. The properties of these matrices as well as their use for the evaluation of the stereo-image are considered. The problem of the adjustment of the stereoscopic image on the 3D LCD wall is also discussed and different types of adjustment are considered. For some of them the corresponding equations are given. The presented approach may be applied also to the case of the multiview autostereoscopic 3D plasma wall. 6055A-27, Session 7 Flatbed-type autostereoscopic display system and its image format for encoding T. Saishu, S. Numazaki, K. Taira, R. Fukushima, A. Morishita, Y. Hirayama, Toshiba Corp. (Japan) We have developed a flatbed-type autostereoscopic display system showing continuous motion parallax as an extended form of a onedimensional integral imaging (1D-II) display system. 1D-II display architecture is suitable for both flatbed and upright configurations because it has a large margin for viewing distance and angle to allow for differences in viewing distances between the near-side and farside edges of the display area. We have also designed an image format specification for encoding 1D-II data. In this parallax image array format, two (or more) viewpoint images whose viewpoint numbers are separated by a constant number are paired, and all of the paired images are combined to obtain an image the same size as the elemental image array. The boundaries inside the paired image correspond to the edges of the 3-D viewing area, while the boundaries between the paired images match the boundaries of DCT blocks and macro blocks. Therefore, 3-D image quality is hardly degraded by lossy codec. The conversion from this format to the elemental image array is simple and does not depend on changes in the viewing distance and associated changes in camera number. Decoding and converting speeds are sufficiently high due to utilization of middleware based on DirectX. 6055A-28, Session 7 Autostereoscopic 3D display A. Schwerdtner, SeeReal Technologies GmbH (Germany) Autostereoscopic 3D Display Autostereoscopic Displays (ASD) tend to become enhanced 2D displays rather than being 3D specific displays. Accordingly, they have to meet strong requirements such as high image quality, large viewing angle, high resolution, etc. These requirements pose enormous challenges on the development of ASDs as their optical design often differs strongly from that of 2D displays. Therefore, most often ASD fulfil only a few features while sacrificing others. High resolution is often retained only with the single user tracking mode and complex opto-mechanical design. Multi user mode displays are easier to manufacture and more cost effective but exhibit lower resolution. The viewing angle is generally limited to 30 through 40 degrees. ASD which are capable of displaying 2D and 3D content most often are limited to switch only between both modes but do not offer simultaneous 2D and 3D displaying.
74
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055A: Stereoscopic Displays and Applications XVII
There are several components contained in ASDs not present in usual displays. Most critical are optical components. Generally beam splitters are needed to focus the left and right stereoscopic images onto the left and right eyes of the observers. SeeReal Technologies focus the images to spots called Sweet Spots (SSP). Any cross-talk between SSPs has to be avoided unless the 3D impression is compromised. This puts hard requirements on optical design and manufacturing. The flat parallel channel optics pose extreme challenges on accurateness. Lateral deviations of the positions of the individual lenses must not exceed 5 microns. The requirements of evenness and thickness are subject to similar conditions. Other issues are waviness and accurateness of the individual lenses. But most importantly, after assembling the optical parts of the ASD the total deviations have to be within a very narrow range depending on the size and resolution of the display. Another issue is backlighting. Due to the optical design, severe attenuations of light intensity may occur. High intensity backlights are needed. SeeReal Technologies achieves high resolution of their ASDs employing tracking. The entire line of tracking from picking-up by the cameras up to the Sweet Spots must not exceed 40 ms. Additional limitations occur when considering ASD as marketable products. Price is an important issue which is strongly affected by features listed above, such as mechanical tracking. The optical components as well as the overall tolerances of the ASD have to be kept reproducibly. The demonstration of a prototype is not sufficient. All components have to be made available. Furthermore, ASDs need to support widespread available software, especially for graphics design such as software based on OpenGL or DirectX. SeeReal Technologies has developed an ASD which it believes to meet most requirements potential customers need. Based on tracking it features native high resolution, switch between 2D and 3D or simultaneous display of 2D and 3D, multi user, and wide angle. As it employs field sequential software almost all 3D applications run on it. Additionally, it has developed the technology for key components such as the high accurate and reproducible beam splitter optics, and tracking. 6055A-29, Session 7 The HoloVizio system T. Balogh, Holografika Kft. (Hungary) We present the HoloVizio system design and give an overview about Holografika's approach to the 3D displaying. The HoloVizio technology uses a specially arranged array of optical modules and a holographic screen. Each point of the holographic screen emits light beams of different color and intensity to the various directions. The light beams generated in the optical modules hit the screen points in various angles and the holographic screen makes the necessary optical transformation to compose these beams into a perfectly continuous 3D view. With proper software control, light beams leaving the pixels propagate in multiple directions, as if they were emitted from the points of 3D objects at fixed spatial locations. We show that the direction selective light emission is a general requirement for every 3D systems and provide quantitative data on the FOV, on the angular resolution, determining field of depth of the displays, affecting the total number of light beams necessary for high-end 3D displaying. We present the results with the 10 Mpixel desktop display and the 50Mpixel large-scale system. We cover the real-time control issues at high pixel-count systems with the HoloVizio software environment and describe concrete 3D applications developed in the frame of European projects.
6055A-30, Session 8 Development of autostereoscopic display system for remote manipulation T. Honda, Y. Kuboshima, K. Iwane, T. Shiina, Chiba Univ. (Japan) When a 3D display system is used for remote operation, the special glasses for looking at the 3D-image disturb the manipulation. So auto-stereoscopic display is preferable for remote manipulation work. However, the eye position area of the auto-stereoscopic display which shows the 3D-image is generally narrow. We constructed a 3D display system which solved these problems. In the system, 1.stereoscopic images displayed on the special LCD are projected on a large concave mirror by a projection lens. 2.The aperture is set between the projection lens and the concave mirror. 3.The real image of the aperture is made at a certain position in vacant space by the concave mirror, and the image position is the viewing zone. By putting both eyes at the position and looking at the concave mirror plane, the viewer can see the stereoscopic image without glasses. To expand the area which can observe the 3D-image, we proposed and constructed the system of the eye-position tracking of the viewing zone by detecting the eye-position of the viewer. A viewer can not only move horizontally and vertically by rotating the concave mirror, but also move to front and back by moving the viewing zone limiting aperture. 6055A-31, Session 8 Ray-space acquisition and reconstruction within cylindrical objective space T. Yendo, T. Fujii, M. Tanimoto, Nagoya Univ. (Japan) A ray-based cylindrical display is proposed that allows multiple viewers to see 3D images from a 360-degree horizontal arc without wearing 3-D glasses. This technique uses a cylindrical parallax barrier and a one-dimensional light source array constructed from such semiconductor light sources as LEDs aligned in a vertical line. The light source array rotates along the inside of the cylindrical parallax barrier, and the intensity of each light is synchronously modulated with the rotation. Since this technique is based on the parallax panoramagram, the density of rays is limited by the diffraction at the parallax barrier. In order to solve this problem, revolving parallax barrier is employed. Two protype displays have been developed and showed high presence 3D image. Especially the newer one is capable of displaying color images whose diameter is 200mm, it is suitable for displaying real object like a human head. We successfully reconstructed a real object within cylindrical space from ray-space data acquired by a video camera rotating around the object. In this paper, we describe details of the system and discuss about ray control method to reconstruct an object from acquired ray-space data. 6055A-32, Session 8 72-directional display having VGA resolution for high-appearance image generation Y. Takaki, T. Dairiki, Tokyo Univ. of Agriculture and Technology (Japan) The high-density directional display, which was originally developed in order to realize a natural three-dimensional display, is not only a three-dimensional display but also a high-appearance display. The appearances of objects, such as glare and transparency, are the results of the reflection and the refraction of rays. The faithful reproduction of such appearances of objects is impossible using conventional two-dimensional displays because rays diffuse on the display screen. The high-density directional display precisely
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
75
Conf. 6055A: Stereoscopic Displays and Applications XVII
controls the horizontal ray directions so that it can reproduce the appearances of objects. The fidelity of the reproduction of object appearances depends on the ray angle sampling pitch. The angle sampling pitch is determined by considering the human eye imaging system. In the present study the high-appearance display which has the resolution of 640Е~400 and emits rays in 72 different horizontal directions with the angle pitch of 0.38° was constructed. Two 72directional displays were combined, each of which consisted of a high-resolution LCD panel (3,840Е~2,400) and a slanted lenticular sheet. Two images produced by two displays were superimposed by a half mirror. A slit array was placed at the focal plane of the lenticular sheet for each display to reduce the horizontal image crosstalk in the combined image. 6055A-33, Session 8 Combining volumetric edge display and multiview display for expression of natural 3D images R. Yasui, I. Matsuda, H. Kakeya, Univ. of Tsukuba (Japan) In the present paper the authors present a novel stereoscopic display method combining volumetric edge display technology and multiview display technology to realize presentation of natural 3D images where the viewers do not suffer from contradiction between binocular convergence and focal accommodation of the eyes. Most of the conventional 3D electronic displays, including multiview displays, use only binocular parallax to make the viewers perceive depth, which causes contradiction between convergence and accommodation of our eyes. This contradiction often causes eyestrain and sickness of the viewer, which are often thought to be the main reason why 3D electronic displays have not been used widely so far. Though several expensive solutions, such as electronic holograms or super-multivew technology, have been invented, these technologies have to process huge amount of data even to show very low resolution images. Therefore it is hard to attain high resolution 3D images with these technologies in the near future. As for cheaper solutions which require less data processing, volumetric display technology is available. 3D images given by volumetric displays do not include contradiction between binocular convergence and focal accommodation. Volumetric displays, however, cannot express occlusion or gloss of the objects. Also volumetric displays require complete volumetric model of the 3D space to be presented, which is hard to capture on the real-time basis in the real world. Thus the conventional volumetric display technology is not a realistic solution for 3D television. To overcome these defects of volumetric displays, the authors propose a method to draw edge areas and flat areas separately. We adopt volumetric display method only for edge drawing, while we adopt stereoscopic approach for flat areas of the image. Since focal accommodation of our eyes is affected only by the edge part of the image, natural focal accommodation can be induced if the edges of the 3D image are drawn on the proper depth. As for the hardware of the volumetric display we can use layered monochrome TFT panels, for we only need to express edges here. Since the conventional stereo-matching technique can give us robust depth values of the pixels which constitute noticeable edges, it is not hard to extract and draw only the edge part of the 3D image on the volumetric display. Also occlusion and gloss of the objects can be roughly expressed with the proposed method since we use stereoscopic approach for the flat area. When we use multiview technology to realize stereoscopy in this system, we can attain a system where many users can view natural 3D objects at the consistent position and posture at the same time. A simple optometric experiment using a refractometer suggests that the proposed method can give us 3-D images without contradiction between binocular convergence and focal accommodation.
6055A-34, Session 8 Adaptive parallax control for multiview stereo panoramas C. Wang, A. A. Sawchuk, Univ. of Southern California Several types of 3D autostereoscopic (AS) visual displays have been developed recently. We are exploring a personal panoramic virtual environment system using stereo panorama capture and AS displays to improve the sense of immersion. The stereo panorama pairs are created by stitching strips that are sampled from images captured with swing panorama structure. We apply Peleg's disparity adjustment algorithm for the generated stereo panorama to achieve large disparity (horizontal parallax) of far away scenes and smaller disparity of closer scenes for stereo perception. Unfortunately, vertical parallax in the stereo panorama still occurs, causing display artifacts and problems in image fusion. To solve these problems, we first present a general image capture model, specify geometrical parameters, and describe the panorama generating process. We then describe an efficient stitching algorithm that matches image regions. The algorithm also corrects for dynamic exposure variation and removes moving objects without manual selection of ground-truth images. We present expressions for the horizontal and vertical parallax, evaluate different parallax measuring techniques, and develop an adaptive vertical and horizontal parallax control algorithm for rendering in different viewing directions. We show several examples subjective tests of stereo panoramas rendered on AS and other stereo displays, and discuss the relative quality of each. 6055A-46, Poster Session Real-time stereographic display of volumetric datasets in radiology X. H. Wang, G. S. Maitz, J. K. Leader, W. F. Good, Univ. of Pittsburgh A workstation for testing the efficacy of stereographic displays for applications in radiology has been developed, and is currently being tested on lung CT exams acquired for lung cancer screening. The system exploits pre-staged rendering to achieve real-time dynamic display of slabs, where slab thickness, axial position, compositing method, brightness and contrast are interactively controlled by viewers. Stereo viewing is by means of shutter-glasses synchronized to a 144 Hz monitor. The system enables viewers to toggle between alternative renderings such as one using distance-weighted ray casting by maximum-intensity-projection, which is optimal for detection of small features in many cases, and ray casting by distance-weighted averaging, for characterizing features once detected. A reporting mechanism is provided which allows viewers to use a stereo cursor to measure and mark the 3D locations of specific features of interest, after which, a pop-up dialog box appears, into which findings can be entered. The system's impact on performance is being tested on chest CT exams for lung cancer screening. Radiologists' subjective assessments have been solicited for other kinds of 3D exams (e.g., breast MRI), their responses have been positive. Objective estimates of changes in performance and efficiency, however, must await the conclusion of our study. 6055A-47, Poster Session Ergonomic evaluation system for stereoscopic video production T. Kawai, S. Kishi, T. Yamazoe, T. Shibata, Waseda Univ. (Japan); T. Inoue, Kanagawa Institute of Technology (Japan); Y. Sakaguchi, K. Okabe, Y. Kuno, Let's Corp. (Japan); T. Kawamoto, Chukyo TV Broadcasting Corp. (Japan)
76
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055A: Stereoscopic Displays and Applications XVII
This paper describes the development of ergonomic evaluation system for producing stereoscopic images. The purpose of the system is to quantify the depth sensation of stereoscopic videos, and evaluate from the view point of safety and amenity. The authors applied the image processing to compute optical flow between right and left videos. The parameters for safety and amenity were examined by referring previous reports and carrying two subjective evaluation experiments. This paper reports the results of the improvement and discusses the validity of the system in terms of stereoscopic video production. 6055A-48, Poster Session Wide-viewing-angle three-dimensional display system using HOE lens array H. Takahashi, H. Fujinami, Osaka City Univ. (Japan); K. Yamada, Hiroshima Institute of Technology (Japan) Integral imaging systems has the problem of the limitation of viewing angle. This paper describes a wide-viewing-angle 3D display system using holographic optical element (HOE) lens array. This display system has a flat HOE lens array. But, the axis of each elemental HOE lens is eccentric. Since every axes of lens is convergent, a flat HOE lens array works as a curved lens array. So, although both a lens array and a screen are flat, this display system has a wide viewing angle. On the other hand, generally, in the integral imaging system each elemental lens has its corresponding area on the display panel. To prevent the image flipping, the elemental image that exceeds the corresponding area is discarded. Therefore, the number of the elemental images is limited and the viewing angle is limited. At the elemental image side, since the axis of each HOE lens is normal, the elemental image does not exceed the corresponding area and the barriers to eliminate the flipped images are not required. Since both a lens array and a screen are flat, the configuration of this display system is simple. 6055A-49, Poster Session Depth maps created from blur information using images with focus at near and at far S. Cho, Electronics and Telecommunications Research Institute (South Korea); J. W. Tam, F. Speranza, R. Renaud, Communications Research Ctr. Canada (Canada); N. Hur, S. Lee, Electronics and Telecommunications Research Institute (South Korea) We demonstrate the use of two original images, near-and farfocused images, with one in which the camera focus is at near and the other at far, respectively, for the generation of depth maps that are useful for DIBR. The method involves the following steps: i) Edges are detected for two original images that are near- and farfocused images, respectively, based on the gradient which is estimated in the luminance function using steerable Gaussian first derivative basis filters. The gradient represents the intensity for each edge using local scale control. ii) Blur is estimated based on the gradient and luminance differential between two regions that are divided by one edge-pixel in order to get an estimate of the blur value which is not unduly influenced by local contrast. iii) Edgedepth map is estimated based on the relation between the blur information of the edges in both near- and far-focused images. vi) Fill-depth map is generated by filling the regions based on the Edgedepth map. v) We generate another viewpoint image to complete a stereoscopic pair by using the depth map and one of the original images. Finally, we evaluate the quality of our depth map and the stereoscopic images.
6055A-50, Poster Session Depth map-based disparity estimation technique using multiview and depth camera G. Um, Electronics and Telecommunications Research Institute (South Korea); S. Kim, K. Kim, Gwangju Institute of Science and Technology (South Korea); N. Hur, Electronics and Telecommunications Research Institute (South Korea); K. Lee, Gwangju Institute of Science and Technology (South Korea) In this paper, we propose a depth map-based disparity estimation technique for the trinocular camera system. Depth map is captured using a depth camera with active sensors, and it is used as an initial estimate of setting adaptive search ranges for solving correspondence problems between image pairs. Existing stereo matching techniques usually produce inaccurate disparities or geometric distortions because of occlusion and mismatching. In case that many objects are widely arranged in 3D space and a baseline between two cameras is long, the search range should be largely set to find all correspondences. But largely fixed search range increases the computation time and the mismatching rate. In our system, we obtained three-view standard definition (SD) images and additional depth map for the center-view image. Since the depth map has relatively reliable accuracy and linearity, the depth map is converted into disparities to be used as reference values for stereo matching using calibrated camera parameters. We perform the stereo matching based on the converted disparity values to increase the matching accuracy and to reduce the computational time. Obtained disparity maps can be used for 3D scene reconstruction and stereoscopic view synthesis using depth image-based rendering (DIBR). 6055A-51, Poster Session A uniform metric for anaglyph calculation Z. Zhang, D. F. McAllister, North Carolina State Univ. We evaluate a new method for computing color anaglyphs based on uniform approximation in CIE color space. The method depends on the spectral distribution properties of the primaries of the monitor and the transmission functions of the filters in the viewing glasses. We will compare the result of this method with several other methods that have been proposed for computing anaglyphs. To compute the color at a given pixel in the anaglyph image requires solving a linear program. We exploit computational properties of the simplex algorithm to reduce computation time by 75 to 80 percent. After computing the color at one pixel, a depth-first search is performed around it to collect all the pixels with color close enough to it. We use criteria for measuring closeness of colors such that the colors of all the collected pixels can be computed using a simple matrix-vector multiplication. We also parallelize the algorithm and implement it on a cluster environment. Many interesting results are described, including the effects of different data dividing schemes. 6055A-53, Poster Session Multiview autostereoscopic display with double-sided reflecting scanning micromirrors A. Nakai, K. Hoshino, K. Matsumoto, I. Shimoyama, The Univ. of Tokyo (Japan) We report on the multiview autostereoscopic display with doublesided reflecting scanning micromirrors. There is a trade-off between the resolution and the number of the view angles in existing multiview stereoscopic displays. In order to solve this problem, we propose the way of projecting time-series pixel data into the discrete
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
77
Conf. 6055A: Stereoscopic Displays and Applications XVII
directions by using scanning micromirrors. As the number of the view angles depends on the number of pixel data projected in one cycle of the scan, the resolution and the number of the view angles can be independently increased. Double-sided reflecting micromirrors actuated by both external magnetic force and Lorentz force were designed and fabricated based on the MEMS (Micro Electro Mechanical Systems) technology. Fabricated micromirrors are 450 µm x 520 µm in size, and characteristics of a micromirror, for example the range of the movement and the resonance frequency, were measured. Then the fabricated micromirrors were integrated with a microlens array, pinhole arrays and an LED matrix to construct a prototype of the multiview autostereoscopic display, and the relationship between the view angle and the light intensity was measured. The validity of our proposed method was proved from the light intensity distribution of this prototype. 6055A-54, Poster Session Depth-enhanced floating display system based on integral imaging J. Kim, Seoul National Univ. (South Korea); S. Min, Information and Communications Univ. (South Korea); Y. Kim, S. Cho, H. Choi, B. Lee, Seoul National Univ. (South Korea) Though the floating display system based on integral imaging can provide moving picture with great feel of depth to the observer, it has limited expressible depth range because the expressible depth range of integral imaging is limited. In this paper, the expressible depth range of the floating display system based on integral imaging is analyzed based on the analysis on the expressible depth range of the integral imaging. Also, a depth-enhanced floating display system based on integral imaging is proposed. In the proposed depthenhanced floating display system, the lens array of the integral imaging is placed at the focal plane of the floating lens. Additionally, the seams on the lens array become less distinct since they are also placed at the focal plane of the floating lens. However, the size of the object changes when the object is out of the overall central depth plane. Thus, the size of objects in elemental image should be rescaled to display correct three-dimensional image. The analysis on the rescaling and the experimental results will be given at the presentation. 6055A-56, Poster Session Three-dimensional sprites for lenticulartype three-dimensional display T. Dairiki, Y. Takaki, Tokyo Univ. of Agriculture and Technology (Japan) The present study proposed the three-dimensional sprite technique that enables to rapidly update three-dimensional images of lenticular type three-dimensional displays. In the present study, the threedimensional sprite was developed for the 72-directional display which consists of a WQUXGA (3,840Е~2,400) LCD panel and a slanted lenticular sheet. It projects a large number of directional images in different horizontal 72 directions with nearly parallel rays. Using a slanted lenticlular sheet, the images interpolation is required in the image interlacing process. The time required for updating the three-dimensional image is about 0.5 second when using a PC (Pentium4 2.4GHz) The three-dimensional sprites were implemented by software. The developed software has the ability to display 40, 12, and 4 sprites at the video rate (30 Hz) for the sprite sizes of 8Е~8, 16Е~16, and 32Е~32, respectively. The three-dimensional sprite technique developed in the present study has following features: (a) three data types (two-dimensional type, threedimensional type, and 360° type), (b) three image size (8Е~8, 16Е~16, and 32Е~32), (c) scaling of sprites depending on zcoordinate, and (d) occlusion of sprites depending on z-coordinate.
6055A-57, Poster Session Optical design considerations for a beam combiner in a StereoMirror (TM) 3D display A. Hochbaum, VAV Consulting; J. L. Fergason, Fergason Patent Properties We present a first order optical model for the StereoMirror(tm) 3D display system consisting of two LCD monitors butted top to top. Each monitor is a source to one image of a stereo pair thus retaining the full native resolution and providing high brightness. A beamcombiner intersecting the angle between the monitors fuses the two images. We present an optical analysis and data regarding the effect of the beam-combiner. The beam-combiner's transmission and reflectivity as well as the local angle of incidence affect the features of the perceived 3D image. We calculate the distribution of angles of incidence at the beam-combiner for various observer positions and points on the monitor and the average AOI for which it should be optimized. We present experimental data of commercial combiners and model results to calculate the optimal combiner parameters that maximize brightness and light efficiency and minimize cross-talk. The polarization directions of 15"-17" monitors are at ± 45° to the plane of incidence leading to an AOI dependent cross-talk effect. We analyze the phenomenon and discuss the magnitude of this effect and means to reduce it to an acceptable level. 6055A-58, Poster Session Horizontal parallax distortion in toe-in camera systems with fisheye lens H. Kang, D. Kim, N. Hur, Electronics and Telecommunications Research Institute (South Korea) In this paper, we propose a novel method that can correct a horizontal parallax distortion (HPD) that happens in toe-in camera configuration with fisheye lenses. In a toe-in camera system, the disparity of the object that is located at convergence point on the convergence plane is 0 where the convergence plane means the specific plane that is located at the distance at which optical axes of two cameras intersect. However, disparities of the other objects that are located the left or right side in the centre on the convergence plane are not zero, and increase according to the distance between objects and convergence point on the horizontal axis. Accordingly, the distribution of disparities is a tangent-like function of the distance from the convergence point to the left or right side on the convergence plane. In order to correct the HPD, we propose a model of a toe-in camera configuration, and calculate the distribution of the HPD. In this paper, we have presented a HPD correction method for a toein camera system with fisheye lenses. In experiment, we used CCD cameras whose focal lengths are 2.5mm and 12mm, known as a fisheye lens, corrected a barrel distortion and HPD from CCD camera, and fixed color inconsistency. 6055A-59, Poster Session Implementation of 3DTV broadcasting system for realistic broadcasting services B. Bae, S. Cho, K. Yun, H. Kang, N. Hur, C. Ahn, Electronics and Telecommunications Research Institute (South Korea) This paper presents a noble 3DTV broadcasting system that is fully compatible with the existing DTV system including MPEG-2 standard, which is a basic one based on various digital broadcasting systems such as ATSC, DVB, OpenCable and so on. The presented system shows the configuration of a flexible, 2D-campatible, and commercially feasible end-to-end 3DTV broadcasting system. To
78
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055A: Stereoscopic Displays and Applications XVII
achieve 2D compatibility, left-view images are encoded by using the MPEG-2 standard and 3D additional information for right-view images is encoded by using the AVC, which is an advanced coding method. Both encoded streams are transmitted in a single program with the proposed map tables. The proposed 3DTV system was verified by two kinds of DTV receivers, which are a commercial and conventional DTV receiver and the developed receiver for receiving 3D information. As results of experiments, we confirmed that the 3D pictures or the 2D pictures can be played on the 3DTV STB according to users' selection and the 2D pictures are only displayed on the commercial DTV receiver without any problems. In the future, the proposed system is expected to be a next-generational one for the advanced digital broadcasting services. 6055A-61, Poster Session Performance analysis of a compact electrooptical 3D adapter with a wide capturing angle S. Kim, J. Lee, E. Kim, Kwangwoon Univ. (South Korea) In this paper, a new 3D adapter system with a lens unit interposed between a photographing lens and an adapter housing for alternately passing right and left moving images of an object there through, wherein the lens unit has an entrance pupil point formed outside the lens unit, the lens unit has a magnification of 1:1, and the lens unit comprises a plurality of symmetrically arranged lenses for reversing the moving images, whereby it is possible to capture moving images with wide picture angles without increasing the size of the adapter housing, and to prevent occurrence of any distortion in the resulting moving images comprised of the integrated right an left images of the object. Some experimental results confirm the superiority of the newly developed 3D adapter system and its usefulness in the practical applications. 6055A-62, Poster Session New method of zoom-convergence interlocked control in the moving parallelaxes style stereoscopic camera J. Lee, S. Nam, J. Lee, C. Park, S. Chung, Korean Broadcasting System (South Korea) We have devised a new and efficient method of zoom-convergence interlocked control in the moving-parallel axes style stereoscopic camera system. As a stereoscopic camera system for broadcasting, the zoom function is indispensable. But, without some special method or apparatus, the zoom operation in the stereoscopic camera system will result in unpredictable change of the disparity in the stereoscopic image. To solve such problems, we devised this new method. First, instead of making the Look-Up-Table by measuring the zoom value and the convergence at each step, we utilized the lens data sheet, which can be obtained from the lens manufacturer, so that we can secure the accuracy and the handiness without any measuring. And, we set up a simple and smart algorithm of our own, which is based on the basic geometry of the stereoscopic camera system. By utilizing the relationship among the inter-camera distance, the object distance and the focal length, namely the distance from lens to image sensor, it can calculate the convergence and keep the convergence or disparity constant at a certain point in spite of zooming operation. From such functions, we can make the stereoscopic contents with smooth zooming and adequate disparity with ease.
6055A-35, Session 9 Integral videography of high-density light field with spherical layout camera array T. Koike, M. Oikawa, N. Kimura, F. Beniyama, T. Moriya, M. Yamasaki, Hitachi, Ltd. (Japan) We propose a spherical layout for a camera array system when shooting images for use in integral videography (IV). IV is an autostereoscopic video image technique based on integral photography (IP) and is one of the preferred autostereoscopic techniques for displaying images. There are many studies on autostereoscopic displays based on this technique indicating its potential advantages. Other camera arrays have been studied, but their purpose addressed other issues, such as acquiring highresolution images, capturing a light field, creating contents for nonIV-based autostereoscopic displays and so on. Moreover, IV displays images with high stereoscopic resolution when objects are displayed close to the display. As a consequence, we have to capture highresolution images in close vicinity to the display. We constructed the spherical layout for the camera array system using 30 cameras arranged in a 6 x 5 array. Each camera had an angular difference of 6 degrees, and we set the cameras to the direction of the sphere center. These cameras can synchronously capture movies. The resolution of the cameras was 640 x 480. With this system, we determined the effectiveness of the proposed layout of cameras and compared image qualities and computational speeds of some algorithms. 6055A-36, Session 9 Imaging properties of microlens arrays for integral imaging system J. Arai, M. Okui, Y. Nojiri, F. Okano, NHK Science & Technical Research Labs. (Japan) Integral imaging (II) system, based on Integral photography proposed by Lippmann, enables three-dimensional (3-D) images corresponding to a viewpoint to be observed under natural light in real time. However, there has been the problem of pseudoscopic images, which must be avoided when creating a real-time system. We propose an elemental optics using a combination of convex lenses in order to avoid pseudoscopic images. Firstly, we describe the structure of an element. For an orthoscopic image, the element of either the capturing or display system must have lateral magnification opposite to the magnification of the usual convex lens. We show that this structure can provide an elemental image for an orthoscopic image. Secondly, overlap between neighboring elemental images must be prevented. We also show that, with our proposed structure, there is no overlap between neighboring elemental images. Finally, we have fabricated a lens array of the proposed elements and have constructed an experimental setup for II system. The results of experiments confirmed that an orthoscopic image is produced in real time. The discussions and experimental results described here show the feasibility of generating highresolution 3-D images in real time using the proposed lens arrays. 6055A-37, Session 9 Comparative study on 3D-2D convertible integral imaging systems B. Lee, H. Choi, J. Kim, Y. Kim, Seoul National Univ. (South Korea) With the development of the display technology, the threedimensional(3D) display attracts much attention as a next generation display technique. In this paper, we compare two types of 3D/2D convertible display methods which we had proposed, and discuss the merits and demerits of each one. One method uses a polymerdispersed liquid crystal(PDLC) and a point light source array. With the adjustment of the PDLC, we can form or eliminate a point light
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
79
Conf. 6055A: Stereoscopic Displays and Applications XVII
source array behind of the spatial light modulator and can switch between the 3D and 2D display modes. The other method uses layered panels. By changing the role of two display panels, the 3D and 2D images can be displayed by the principles of the integral imaging and liquid crystal display(LCD) respectively. Both systems require no mechanical movement and can provide a full color 3D image with continuous viewpoints and no special glasses. The one using a PDLC is advantageous for the depth of displayed 3D image while the other one is advantageous for the resolution. We also discuss the generation of the elemental images and some improved methods for both two types of 3D/2D convertible displays. A method to prevent the occlusion problem will also be proposed. 6055A-39, Session 10 Application of 3DHiVision: a system with a new 3D HD renderer P. Sun, Sun Advanced Engineering, Inc. (Japan); S. Nagata, InterVision (Japan) This paper discusses about some strategies that will help to solve the problems that prevent 3D Stereo from being popular. A 3DHV (3DHiVision) System Solution Modern projection systems and stereo LCD panels have made it possible to enable a 3D stereo video experience for many more people in a broader range of applications. The key limitations to more mainstream applications of 3D video have been the availability of 3D contents and the cost and complexity of 3D video production, content management and playback systems. Even with the modern PC based video production tools, projection systems and increased interest in 3D applications, the 3D video industry remains small and stagnant. The inescapable fact is that the production and playback of high quality 3D video is still expensive. The system has accepted the challenge of overcoming these difficulties and has created a complete end-toend 3D Video system based on an embedded PC platform, which significantly reduces the cost and complexity of creating museum quality 3D video. Amateur or professional film makers will be able to easily create, distribute and playback 3D video contents. 6055A-40, Poster Session 3D animation in three dimensions: the rocky road to the obvious H. Murray, IMAX Corp. (Canada) That animation created using CG modeling and animation tools is inherently three-dimensional is well known. In the middle to late nineties IMAX Corporation began actively exploring CG animated features as a possible source of economically viable content for its rapidly growing network of stereoscopic IMAX(r) 3D theatres. The journey from there to the spectacular success of the IMAX(r) 3D version of The Polar Express is an interesting mix of technical, creative and production challenges. For example 3D animations often have 2D elements and include many sequences that have framing, composition and lens choices that a stereographer would have avoided had 3D been part of the recipe at the outset. And of course the decision to ask for a second set of deliverables from an already stressed production takes nerve. The talk will cover several of these issues and explain why the unique viewing experience enabled by the wide-angle geometry of IMAX(r) 3D theatres makes it worth all the pain.
6055A-41, Session 12 A method of real-time construction of full parallax light field K. Tanaka, S. Aoki, Sony Corp. (Japan) We designed and implemented a dynamic light field acquisition and reproduction system, which serves as a 3D live video system for multiple viewers. The acquisition unit consists of circularly arranged NTSC cameras surrounding an object. The display consists of circularly arranged projectors and a rotating screen. The projectors are constantly projecting images captured by the corresponding cameras onto the screen. The screen rotates around an in-plane vertical axis so that it faces each of the projectors in sequence. Since the surfaces of the screens are covered by light-collimating plastic films with their louver patterns being vertical, which have a role to make the screen retro-reflective horizontally, a viewer can observe only an image projected by a projector located in the same direction as the viewer. Thus the viewer can see a dynamic image of an object according to his or her head position. We evaluated the system by projecting both objects and human figures, and confirmed that the system can reproduce live video sequences with horizontal parallax. Application areas of this system include product design review, sales promotion, art exhibition, fashion show, and sports training with form checking. 6055A-43, Session 12 Simulation of 3D image depth perception in a 3D display using two stereoscopic displays at different depths K. Uehira, Kanagawa Institute of Technology (Japan) We studied a new 3-D display that uses two stereoscopic displays instead of two 2-D displays in a depth-fused 3D display. We found that two 3-D images with the same shape displayed at different depths by the two stereoscopic displays were fused into one 3-D image when they were viewed as overlapping. Moreover, we found that the perceived depth of the fused 3-D image depends on both the luminance ratio of the two 3-D images and their original perceived depth. This paper presents the simulation results for the perceived depth of the fused 3-D image on the new 3-D display. We applied a model in which the human visual system uses a low-pass filter to perceive the fused image, the same as on a conventional DFD display. The simulation results revealed that the perceived depth of the fused image changed depending on both the luminance ratio of the two 3-D images and their original perceived depth, as in the subjective test results, and the low-pass filter model accurately presented the perception of a 3-D image on our 3-D display. 6055A-44, Session 12 Innovative stereoscopic display using variable polarized angle J. E. Gaudreau, PolarScreens, Inc. (Canada) and MacNaughton, Inc.; M. Bechamp, PolarScreens, Inc. (Canada); B. MacNaughton, V. S. Power, MacNaughton, Inc. PolarScreens has developed an innovative stereoscopic display (US patent # 5,629,798) that has the unique advantage of displaying the two images without any multiplexing. PolarScreens uses 2 LCD panels stacked on one another where the second LCD adds extra information to the photon using a polar coordinate transformation algorithm. The first LCD controls total pixel intensity and the second controls left-eye/right-eye distribution ratio. Notably, this is the only technology where one photon contains information for both eyes! Based on this theory, PolarScreens has developed the technologies required to make it practicable and built a 19in, 1280x1024 stereoscopic desktop monitor with very low crosstalk, high
80
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055A: Stereoscopic Displays and Applications XVII brightness and awide angle of vision using passive polarized glasses. The biggest problem in stereoscopic monitors is crosstalk. It is a measure of the interference between the left and right pictures. Existing stereoscopic monitors produce around 10% crosstalk. This monitor brings this down to 2%. That's a 500% improvement. Moreover, at the push of a button, the display can switch between different modes of operation: · Stereoscopic (user wear passive 3D glasses) · Normal 2D display · Contrast Enhanced 2D (just by adding a film on the display surface) · Private (where only the person wearing special glasses can see the image). This technology, as show with the first product manufactured by NuVision, has a great potential in the professional as well as entertainment markets. 6055A-45, Session 12 A novel walk-through 3D display S. DiVerdi, A. Olwal, I. K. Rakkolainen, T. Hцllerer, Univ. of California/Santa Barbara We present a novel walk-through 3D display based on the patented FogScreen, an "immaterial" indoor 2D projection screen, which enables high-quality projected images in free space. We extend the basic 2D FogScreen setup in three major ways: First, we implemented various techniques for tracking a viewer's head, enabling correct perspective 3D projection for one viewer, second, we added support for stereoscopic imagery, and third, we present the front and back views of the graph¬ics content on the two sides of the FogScreen, so that the viewer can cross the screen to see the content from the back. The 3D objects in free space look natural when viewed on-axis. While the image quality is not perfect from off-axis viewing angles, our system produces a believable 3D impression, implementing the first human-scale reach- and walk-through pseudo-volumetric display. Our informal user evaluation of the system suggests that the display leads to an enhanced viewing experience. It creates a strong visual effect of 3D objects floating in air, even in the case when the image is not stereoscopic. This is a first step in the direction of a truly volumetric walk-through display.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
81
Conf. 6055B: The Engineering Reality of Virtual Reality 2006 Thursday 19 January 2006 Part of Proceedings of SPIE Vol. 6055 Stereoscopic Displays and Virtual Reality Systems XIII
6055B-63, Session 13 Texturing of continuous LoD meshes with the hierarchical texture atlas H. Birkholz, Univ. Rostock (Germany) For the rendering of detailed virtual environments, trade-offs have to be made between image quality and rendering time. An immersive experience of virtual reality always demands high frame-rates with the best reachable image quality. Continuous Level of Detail (cLoD) triangle-meshes provide an continuous spectrum of detail for a triangle mesh that can be used to create view-dependent approximations of the environment in real-time. This enables the rendering with a constant number of triangles and thus with constant frame-rates. Normally the construction of such cLoD mesh representations leads to the loss of all texture information of the original mesh. To overcome this problem, a parameter domain can be created, in order to map the surface properties (colour, texture, normal) to it. This parameter domain can be used to map the surface properties back to arbitrary approximations of the original mesh. The parameter domain is often a simplified version of the mesh to be parameterised. This limits the reachable simplification to the domain mesh which has to map the surface of the original mesh with the least possible stretch. In this paper, a hierarchical domain mesh is presented, that scales between very coarse domain meshes and good property-mapping. 6055B-64, Session 13 Optimal approximation of head-related transfer function's zero-pole model based on genetic algorithm J. Zhang, Southeast Univ. (China) In the research of spatial hearing and virtual auditory space, it is important to effectively model the head-related transfer functions (HRTFs). Based on the analysis of HRTFs' spectrum and some perspectives of psychoacoustics, this paper applied multiple demes' parallel and real-valued coding genetic algorithm (GA) to the approximation of HRTFs' zero-pole model. Using the logarithmic magnitude's error criterion for human auditory sense, the results show that the performance of GA is averagely 39% better than that of traditional Prony method, and 46% than that of Yule-Walker algorithm. 6055B-65, Session 13 Multiprojector image distortion correction scheme for curved screens on the example of the Cybersphere B. V. Shulgin, J. Ye, V. H. Raja, Univ. of Warwick (United Kingdom) A problem appearing in the virtual reality (VR) is the image distortion and blending for curved screens. Because of the different distance from the part of a curved screen to projectors the different parts of the projected image have different magnification that leads to distortion. There are ways to solve this problem via a special circuit implementation within the image projectors or via special image correction PC based boxes. The projected image should be predistorted initially in order to get an undistorted image on the curved screen. These existing solutions are limited in number of projectors or screen configurations. We propose an original image correction algorithm which can work on screen surfaces of arbitrary geometry. It was tested on the Cybersphere, the VR technology developed by the University of Warwick. The method allows blending images from
multiple projectors confining the corrected image within given boundaries. We propose software implementation of the algorithm which allows using it for any programs not limited by OpenGL technology. We also propose using it for distributed image rendering and projection such as Chromium based sets. We apply the method for variable curvature screen, a novel patented VR set developing at the Warwick Manufacturing Group, University of Warwick. 6055B-66, Session 13 3D workflow for HDR image capture of projection systems and objects for CAVE virtual environments authoring with wireless touch-sensitive devices M. J. Prusten, Optical Design Labs.; M. K. McIntyre, Total Eclipse Studios; M. Landis, The Univ. of Arizona We present a 3D workflow pipeline for the high dynamic range (HDR) image capture of projected scenes or objects, for authoring in CAVE virtual environments. The methods of high dynamic range digital photography of environments vs. objects are reviewed. The image collection of a scene or object requires a digital photographic procedure and post software libraries to reconstruct the high dynamic photographs for the specific virtual reality applications. Once the high dynamic range images have been acquired our workflow pipeline, CAVEPIPE, can be incorporated. Samples will be reviewed of both types of virtual authoring being, the actual CAVE environment and a sculpture. Our virtual reality environment, from FakeSpace, consists of four 8x10 ft screens on three sides and a floor, which are the rear projection screens and the boundaries of the 3D space. The projected images are edge-blended multiple projector tiles with 1600x1200 resolution at 96 Hz refresh. The software library tools being used at the AZ-LIVE (Arizona Laboratories for Immersion Visualization Environments) are a series of software tools incorporated into a pipeline called CAVEPIPE. This pipeline includes: OpenGL performer, CAVELib, NUKE and the major 3D animation packages, Lightwave, Maya, Softimage, and Houdini. OpenGL Performer is a high performance 3D rendering toolkit for developing real-time, multi-processed, interactive graphics. The CAVELib routines in the virtual reality applications provide the building blocks for the multiple display devices, the viewer-centered perspective camera projections, stereoscopic viewing, and the 3D tracked input devices. With the captured data rendered, we composite the various image layers into image sequences. Both object and scene images can be composited into a photorealistic scene. The NUKE compositor is a new tool that supports object importing, projection mapping, and 3D camera import from all industry standard trackers. This allows for high-resolution objects and scenes to be composted together in natural illumination environments and presented in our CAVE virtual reality environment. This workflow for authoring 3D content has been demonstrated in: atomic modeling, architectural modeling, and on HDR captured data of virtual projection environments and HDR images of objects. The 3D object and scene importers are being expanded to handle data formats from Houdini, Softimage XSI, Maya animation packages and the NUKE compositor. The 3D scene acquisition are being updated acquire data not just from a fisheye lens, but also from a wide-angle lens to eliminate the aberrations at the edge of field. The use of touch panel display systems and wireless devices are also being customized and added to the CAVE interface. This will allow the user to have more useful graphical user interfaces (GUI) while exploring and developing virtual reality content in the CAVE.
82
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055B: The Engineering Reality of Virtual Reality 2006
6055B-67, Session 14 Examination of corner vane estrangement evaluation method for a circular tunnel H. Yokoyama, O. Fujishima, Hitachi, Ltd. (Japan) For the next several years, tunnel construction several with highway construction are planned around a Japanese city. Furthermore, several ventilation facilities constructions are planned by these constructions. There are circle or rectangle tunnel, and rectangle duct in these tunnel ventilation facilities. And a rectification board called a corner vane is installed to reduce a pressure loss on these corner parts. For design of conventional corner vane, 2D CAD has been generally used. However, the design in curved surface part in ceiling part of shield tunnel became difficult. Because, the ceiling part is complicated curved surface shape. Furthermore, When a difference (a building error) of form of actual form was identified as a design value, Estrangement occurs between a ceiling part and corner vane ends. Therefore, in a design or installing, correspondence method had to be examined. In this problem, when laser scanner that can acquire wide 3D data in real time was used for 3D measurement of ceiling side of circle tunnel, an actual form will be grasped. Furthermore, an actual form can reappear on 3D CAD by taking measurement result. This point was paid attention by the authors; laser scanner was made to grapple together with 3D CAD, The estrangement of a corner vane for a circular tunnel-related evaluation method and effectiveness were examined. Authors will be content about this topic. 6055B-68, Session 14 Virtual technical support for field engineers in the water and ventilation hygiene industry I. A. Nicholas, Cardiff Univ. (United Kingdom); D. Kim, Aqua Marc Ltd. (United Kingdom) This paper presents an industrial application of VR, which has been integrated as a core component of a virtual technical support system. The problems that often impede conventional technical support and the way in which a virtual technical support system can overcome them are discussed. Field engineers are able to use the system to access improved information and knowledge through their laptop computers while on-the-job. Thereby, taking advantage of scarce investment resources. When used in synergy, the application of VR, multimedia, coordinated multiple views and knowledge-based technologies can be shown to significantly reduce technical support costs. Initial results are presented showing the effectiveness and benefits of such a system for field engineer support in the water and ventilation hygiene industry. 6055B-69, Session 14 Virtual reality in construction industry: a requirement compatibility analysis approach J. Ye, B. V. Shulgin, V. H. Raja, Univ. of Warwick (United Kingdom) Virtual Reality (VR) is regarded as a high-end user-computer interface that involves real-time simulation and interactions through multiple sensorial channels. It is assumed that VR will reshape the interaction interfaces between user and computer technology by offering new approaches for the communication of information, the
visualisation of processes and the creative expression of ideas. The VR application in construction has a relatively long history but its successful stories are not heard quite often. In this paper, the authors have explored how much further the construction industry could be supported by new three dimensional (3D) VR technologies in different construction processes. The design information in the construction industry has been discussed first followed by a detail construction process analysis. A questionnaire survey has been conducted and the results of the survey are presented and discussed. As an investigation into the application of 3D VR technologies in the context of the construction processes, the benefits and challenges of current and potential applications of 3D VR in the construction industry have been identified. This study also reveals the strengths and weaknesses of 3D VR technology applications in the construction processes. Suggestions and future works are also provided in this paper. 6055B-70, Session 14 Adding tactile realism to a virtual reality laparoscopic surgical simulator with a costeffective human interface device I. W. Mack, Queen's Univ. Belfast (United Kingdom); S. Potts, The Royal Group of Hospitals (United Kingdom); K. R. McMenemy, R. S. Ferguson, Queen's Univ. Belfast (United Kingdom) The laparoscopic technique for performing abdominal surgery requires a very high degree of skill in the medical practitioner. Much interest has been focused on using computer graphics to provide simulators for training surgeons. Unfortunately, these tend to be complex and have a very high cost, which limits availability and restricts the length of time over which an individual can practice their skills. There is no reason why the cost should be high as computer game technology delivers excellent graphics that easily fulfil any requirement of a surgical simulator. If it was down to the graphics alone every surgeon could have a training and practising tool on their laptop. It is the appropriate human interface hardware, the equivalent of the `joystick', which is missing. This paper presents a design for a very low cost device to address this vital issue. The design encompasses: the mechanical construction, the electronic interfaces and the software protocols to mimic a laparoscopic surgical set-up. Thus the surgeon has the capability of practising two-handed procedures with the possibility of force feedback. The force feedback and collision detection algorithms allow surgeons to practice realistic operating theatre procedures with a good degree of authenticity. 6055B-71, Session 15 Inverse perspective M. Dolinsky, Indiana Univ. This paper will discuss the potentiality towards a methodology for creating perceptual shifts in virtual reality (VR) environments. A perceptual shift is a cognitive recognition of having experienced something extra-marginal, on the boundaries of normal awareness, outside of conditioned attenuation. Perceptual shifts are further defined, demonstrated in a historical tradition, analyzed through various categories of sensory illusions and explained via biological perceptual mechanisms and the sciences, including neuroscience and cognitive science. This paper explores perspective, illusion and projections to situate an artistic process in terms of perceptual shifts. Most VR environments rely on a single perceptual shift while there remains enormous potential for perceptual shifts in VR. Examples of paintings and VR environments that develop this idea will be presented.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
83
Conf. 6055B: The Engineering Reality of Virtual Reality 2006
6055B-72, Session 15 Virtual reality and the unfolding of higher dimensions J. C. Aguilera, Univ. of Illinois at Chicago As augmented reality evolves, the need for spaces that are responsive to structures independent from three dimensional spatial constraints, become apparent. The visual medium of computer graphics may also challenge these self imposed constraints. If one can get used to how projections affect objects in three dimensions, it may also be possible to compose a situation in which to get used to the variations that occur while moving through higher dimensions. The presented application is an enveloping landscape of concave and convex forms which are determined by the orientation and displacement of the user in relation to a grid made of hypercubes. The interface accepts input from tridimensional and four dimensional transformations, and smoothly displays such interactions in realtime. The motion of the user becomes the graphic element whereas the higher dimensional grid references to his/her position relative to it. The user learns how motion inputs affect the grid, recognizing a correlation between the input and the transformations. Mapping information to complex grids in virtual reality is valuable for engineers, artists and users in general because navigation can be internalized like a dance pattern, and further engage us to maneuver space in order to know and experience. 6055B-73, Session 15 Framing the magic D. Tsoupikova, Univ. of Illinois at Chicago This paper will explore how the aesthetics of the virtual world affects, transforms, and enhances an immersive emotional experience of the user. What we see and what we do upon entering the virtual environment influences our feelings, mental state, physiological changes and sensibility. To create a unique virtual experience the important component to design is the beauty of the virtual world based on the aesthetics of the graphical objects such as textures, models, animation, and special effects. When the user is immersed into the visual art world, what they see inspires imagination, remembrance and positive feelings. Through aesthetics the immersive effect is much stronger and more compelling. In order to control a user emotions the virtual reality environment must be customized for an individual experience. But quality of the graphics, poetics and the aesthetics of the virtual world can be made to evoke certain sensible changes in the user's state of mind such as engaging interactivity, fun, thought provocation, and challenging storyline. These changes and effects on the user can be adopted from the art and science by researching the fields of color theory, art education, art therapy, visual music, design, architecture and art history. Many artists these days work on the aesthetics of virtual reality, based upon their experiences in traditional arts such as painting, sculpture, design, education and music to create higher quality stereo graphics of virtual worlds. This paper will use examples of virtual art projects such as World Skin by Maurice Benayoun. Beat Box by Margaret Dolinsky, PAAPAB by Dave E. Pape, and Josephine Anstey, Yiggdrasil by Bino and Cool, Uzume by Petra Gemeinboeck, Osmose by Char Davies and others to compare the immersive power of aesthetics into their emotional effect and how stereo technology could benefit the creativity. Arthur Clarke, the renowned science fiction writer once said that sufficiently complex technology was indistinguishable from magic. The ratio between virtual reality aesthetics and human perceptions is the key for the development of the impressive immersive experience.
6055B-74, Session 15 Virtual reality, immersion, and the unforgettable experience J. F. Morie, Univ. of Southern California Thirty years ago, working in the nascent field of virtual reality meant combining simple polygons, adding some textures and figuring out a method for navigating that space. Today, we think not only about the space, but also time and story, the amount of agency provided, the expected degrees of immersion and the quality of presence. There are new flavors of virtual reality, too, from the private, meditative ones to massively multiplayer online versions. There are virtual worlds, virtual environments, virtual stories, interactive plays and collaborative performances. This expansion of VR territory requires a more sophisticated approach to understanding exactly what VR is, and ultimately, what it still might be. At the core of virtual reality is the desire to make a technological work that provides a participant with an unforgettable experience. Diverse disciplines, such as art, cognitive science, psychology, phenomenology, embodiment, performance, and play all contribute to the design of such experiences. The distinction between spatial and social immersion is especially important for categorizing modern virtual realties and this paper will explore how these two approaches contribute uniquely to the design of the ultimate "unforgettable experience". 6055B-75, Session 16 Teleoperation interface for mobile robot with perspective-transformed virtual 3D screen on PC display T. Kimura, H. Kakeya, Univ. of Tsukuba (Japan) The authors propose an inexpensive human interface for teleoperation of mobile robots by giving a perspective-transformed image of a virtual 3D screen on a standard PC display. Conventional teleoperation systems of mobile robots have used multiple screens for multiple cameras or a curved screen for a wide view camera, both of which are expensive solutions intended only for professional use. We adopt a single standard PC display as the display system for the operator to make the system affordable to all PC users. To make the angular location perceivable with a 2D display, the authors propose a method to show on the flat screen a perspectivetransformed image of a virtual 180-degree cylindrical screen. In this system the image shown on the 2D screen preserves angular information of the remote place, which can help the operator grasp the angular location of the objects in the image. The result of the experiments indicates that the perspective-transformed images of the cylindrical screen can give the operator a better understanding of the remote world, which enables easier and more instinctive teleoperation. 6055B-76, Session 16 An orientation sensing interface for portable situation awareness displays J. Bleecker, Univ. of Southern California No abstract available
84
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6055B: The Engineering Reality of Virtual Reality 2006
6055B-77, Session 16 An interactive camera placement and visibility simulator for image-based VR applications A. State, G. Welch, A. Ilie, The Univ. of North Carolina at Chapel Hill Many VR-based or surveillance-oriented applications use multicamera setups in view of 3D reconstruction or observation of target scenes and environments. Choosing the number of cameras as well as their locations, orientations, fields of view and other parameters for such setups is a difficult task. This paper describes an interactive software simulator that assists in the development of such scenarios. The simulator allows interactive placement and manipulation of multiple cameras within a pre-modeled 3D environment. It shows the user the exact scene coverage for each camera, by projecting user-specified 2D patterns from each camera onto the scene, taking into account occluded areas. The user can navigate through the modeled scene, investigate lacking or overlapping coverage and interactively adjust camera parameters. The 2D patterns can be structured to indicate local imaging resolution at each point in the scene. Specialized patterns and projection techniques can be used to describe the dynamic coverage areas of pan-tilt-zoom cameras. We describe the simulator and its interface and show an example multi-camera setup for remote 3D medical consultation including some preliminary results. 6055B-78, Session 16 Overview of virtual camera mechanisms for collaborative virtual environments: an application to the VRIMOR project E. E. Alvarez, A. A. De Antonio, Univ. Politйcnica de Madrid (Spain) This article presents a research on the control of automated cameras in dynamic 3D virtual environments, which analyses the different branches that are currently being developed and improved in relation to the placement and movement of the virtual cameras. The aim of this research is to choose a method to design a management system for automatic cameras to be applied in the project VRIMOR. This project allows operators of nuclear power plants to design interventions with the aid of virtual mannequins with humanoid form. The designed intervention will be used for the learning of the operators in collaborative virtual environments. It is at this stage when the automated management of virtual cameras becomes relevant. 6055B-79, Session 16 Synthecology: `sound' use of audio in teleimmersion G. A. Baum, SUNY/Univ. at Buffalo; M. Gotsis, Univ. of Southern California; B. Chang, R. Drinkwater, D. St. Clair, Art Institute of Chicago We examine historical audio applications used to provide real-time immersive sound for CAVE environments and discusses their relative strengths and weaknesses. We examine and explain issues of providing spatialized sound immersion in real-time virtual environments, some problems with currently used sound servers, and a set of requirements for an `ideal' sound server. We present the initial configuration of a new cross-platform sound server solution using open source software and the Open Sound Control (OSC) specification for the creation of real-time spatialized audio with CAVE applications, specifically ygdrasil environments. The application, Another Sound Server,establishes an application
interface (API) using OSC, a logical server layer implemented in Python, and an audio engine using SuperCollider. We discuss spatialization implementation and other features. Finally, we document the Synthecology project which premiered at WIRED Nextfest 2005 and was the first virtual environment to use Another Sound Server. We also discuss various techniques that enhance presence in networked virtual environments, as well as possible and planned extensions of Another Sound Server.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
85
Conf. 6056: Three-Dimensional Image Capture and Applications VI Monday-Tuesday 16-17 January 2006 Part of Proceedings of SPIE Vol. 6056 Three-Dimensional Image Capture and Applications VII
6056-01, Session 1 A novel design of grating projecting system for 3D reconstruction of wafer bumps Y. Shu, Xi'an Jiaotong Univ. (China); R. C. Chung, J. Cheng, The Chinese Univ. of Hong Kong (Hong Kong China); E. Y. Lam, The Univ. of Hong Kong (Hong Kong China); K. S. M. Fung, F. Wang, ASM Assembly Automation Ltd. (Hong Kong China) A challenge in the semiconductor industry is the 3D inspection of solder bumps grown on wafers for direct die-to-die bonding. In an earlier work we proposed a mechanism for reconstructing wafer bump surface in 3D, which is based upon projecting a binary grating to the surface from an inclined angle. For the purpose of 3D reconstruction with high speed and accuracy, the requirements for the projection lens system are as the followings: (1) having a tilted angle, usually about 60 degree, between the projection plane and the optical axis; (2) having high bandwidth to let high-spatialfrequency harmonics contained in the binary grating pass through the lens and be projected onto the inspected surface properly; (3) having high Modulation Transfer Function (MTF); (4) having large Field of View (FOV); and (5) having large Depth of Filed (DOF) that corresponds to the depth range or height of the inspected surface. The above requirements lead to many difficulties in the design of the projection lens system. We designed a system consisting of a grating and several pieces of spherical lens, that addresses the requirements. To reduce the lens aberrations, the grating is laid out with a tilting angle specifically to make the grating, the lens, and the image plane intersect at the same line. Such a system can project a high spatial-frequency binary grating onto the inspected surface properly. Simulation results, including performance analysis and tolerance analysis, are shown to demonstrate the feasibility of the design. 6056-02, Session 1 Measurement of discontinuities on 3D objects using digital moirй J. Liao, L. Cai, The Hong Kong Univ. of Science and Technology (Hong Kong China) In this paper, a two-dimensional, binary fringe pattern is designed as structured light for 3D measurement. A feature, i.e., a white cross, is placed in the center of the fringe grating. The cross serves as axes of a reference frame. White square grids are alternated by black stripes vertically or horizontally elsewhere. Relative position of a given point on the fringe with respect to the center can be identified. When the fringe pattern is projected onto the surface of the object, its image is distorted. Therefore, image processing and pattern recognition algorithms are designed to calculate which row and column the particular point belongs to in the original fringe-frame. The pair of emitting and receiving angles for each point in the fringe and CCD frames, respectively, is acquired. And coordinate of each 3D point can be calculated. Compared with traditional digital moirй methods, this method achieves an absolute measurement of 3D surface because the information contained in the pattern is globally structured. Therefore, discontinuity measurement can be solved more easily. And resolution of the proposed method is larger than that of current methods of coding patterns under the same line width limitation due to principle of pattern design.
6056-03, Session 1 High-speed and high-sensitive demodulation pixel for 3D imaging B. Bьttgen, T. Oggier, Ctr. Suisse d'Electronique et de Microtechnique SA (Switzerland); P. Seitz, Swissnex; F. Lustenberger, Ctr. Suisse d'Electronique et de Microtechnique SA (Switzerland) Optical time-of-flight (TOF) distance measurements can be performed using so-called smart lock-in pixels. By sampling the optical signal 2, 4 or n times in each pixel synchronously with the modulation frequency, the phase delay between the emitted and reflected signal is extracted and allows determining the object distance. The high integration-level of such lock-in pixels enables the real-time acquisition of the three-dimensional environment without using any moving mechanical components. A novel design of the 2-tap lock-in pixel in a 0.6µm semiconductor technology is presented. The pixel was implemented on a sensor with QCIF resolution. Although a small pixel size of 40x40um^2 has been realized, high optical sensitivity is still achieved due to an optical fill factor of larger than 20% and high quantum efficiency. The optimized buried channel allows high-speed operation of the device resulting in a near-optimum demodulation performance and precise distance measurements which are almost exclusively limited by photon shot noise. In-pixel background-light suppression allows the sensor to be operated in an outdoor environment with full sun-light incidence. The highly complex pixel functionality of the sensor was successfully demonstrated on the new SwissRanger SR3000 3DTOF camera design. 6056-04, Session 1 A QVGA-size CMOS time-of-flight range image sensor with background light charge draining structure T. Ushinaga, I. Abdul Halin, T. Sawada, S. Kawahito, Shizuoka Univ. (Japan); M. Homma, Sharp Corp. (Japan); Y. Maeda, Suzuki Motor Corp. (Japan) Range image sensors can be used in a variety of applications such as in automobile, medicine, robot vision systems, security and so on. Time-of-Flight (TOF) range imaging is a 3D image capture method where the roundtrip time of flight of light pulses from a sensor to a target is sensed. Range is then determined by multiplying half of the TOF with the speed of light constant, c. Recently TOF sensors are implemented using CCD, CCD-CMOS hybrid, and CMOS technologies. However the reported maximum spatial resolution is only 160 Е~ 124. This paper presents a new type of CMOS TOF range image sensor using single layer polysilicon gates on thick field oxide. A structure for background light induced charge reduction is also included in the pixel. The sensor with a spatial resolution of 336 Е~ 252 (QVGA) pixels of 15 Е~ 15µm(c)~ size was fabricated in a 0.35µm standard CMOS process. For preventing photoelectrons to be captured at SiSiO2 interface, an additional process step for buried channel structure is used. The light source used is an array of near-infrared LEDs. Using a small duty cycle (1-10%) light pulse and the charge draining structure drastically reduces the effect of background illumination, because reducing the duty cycle can increase tolerable maximum optical power in the LED. The sensor achieves a minimum range resolution of 2.8cm at a framerate of 30fps, while the resolution is improved to 4.2mm at 3fps, corresponding to 10 frames averaging.
86
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6056: Three-Dimensional Image Capture and Applications VI
6056-05, Session 2 Overview of 3D surface digitization technologies in Europe N. D'Apuzzo, Homometrica Consulting (Switzerland) The solutions for 3D surface measurement offered by major European companies can be divided into different groups depending on various characteristics, such as, technology (e.g. laser scanning, white light projection), system construction (e.g. fix, on CNC/robot/ arm) or measurement type (e.g. surface scanning, profile scanning) Crossing between the categories is possible, however, the majority of commercial products can be divided into the following six groups: (a) laser profiler mounted on robot/CNC/arm, (b) white light projection system mounted on robot and/or object on rotating platform, (c) laser point measurement system where both sensor and object move, (d) hand held profiler or point measurement systems, (e) portable laser scanning or white light projection surface measurement systems, (f) dedicated systems. This paper presents an overview of the different 3D surface digitization technologies commercially available in the European market. It describes them with their advantages and disadvantages. Various examples of their use are shown for different application fields. A special interest is given to applications regarding the 3D surface measurement of the human body. 6056-06, Session 2 Automatic 3D real world surface texture mapping using perspective projection methods C. Shih, MingDao Univ. (Taiwan) Traditional digital reconstruction methods require large amount of human interactions that will impede the speed and accuracy of full detail full color model reconstruction. This research proposes to explore the techniques to build an automatic procedure that can correlate a 3D mesh of a CAD model with a set of photos taken from the corresponding real world object. Correct pixel mapping from 2D space to 3D coordinate is the key to the success of such system. Moreover due to large amount of human intervention during picture correlation step, texture wrapping deviation is inevitable and left unexplored. A research on the factors affecting the precision of perspective photo mapping and texturing seems to lighten up a way leading to automatic and accurate full color mesh model display. The research plan to use the following steps in achieving the goal. (1)build an initial 3D image coordinate system by correlating a few pictures taken on the object. These pictures will cover a few limited number of carefully selected control points located on the real object. (2)derive a composite transformation matrix that combines the image coordinate system with the local mesh coordinate system. (3)use a visibility checking and Z buffer algorithm to do view planning and generate a few photo angles that will allow complete coverage of all the meshes. (4)for every photo taken, a perspective photo mapping procedure is taken to map corresponding pixel to the 3D mesh. The results of this research can directly reduce human interaction and increase the productivity of digital recovery effort. Not only does it increase the value of digital reconstruction industry, but also prove the effectiveness of iterative least square fit theory on texture mapping.
6056-07, Session 2 Virtual confocal macroscopy P. M. Hanna, U.S. Air Force; B. D. Rigling, Wright State Univ. There is a need for persistent-surveillance assets to capture highresolution, three-dimensional data for use in assisted target recognizing systems. Passive electro-optic imaging systems are presently limited by their ability to provide only 2-D measurements. We describe a methodology and system that uses existing technology to obtain 3-D information from disparate 2-D observations. This data can then be used to locate and classify objects under obscurations and noise. We propose a novel methodology for 3-D object reconstruction through use of established con-focal microscopy techniques. A moving airborne sensing platform captures a sequence of georeferenced, electro-optic images. Con-focal processing of this data can synthesize a large virtual lens with an extremely sharp (small) depth of focus, thus yielding a highly discriminating 3-D data collection capability based on 2-D imagery. This allows existing assets to be used to obtain high-quality 3-D data (due to the fine zresolution). This paper presents a stochastic algorithm for reconstruction of a 3D target from a sequence of affine projections. We iteratively gather 2-D images over a known path, detect target edges, and aggregate the edges in 3-D space. In the final step, an expectation is computed resulting in an estimate of the target structure. 6056-08, Session 2 A robust algorithm for estimation of depth map for 3D shape recovery A. Malik, T. Choi, Gwangju Institute of Science and Technology (South Korea) Three-dimensional shape recovery from one or multiple observations is a challenging problem of computer vision. In this paper, we present a new focus measure for calculation of depth map. That depth map can further be used in techniques and algorithms leading to recovery of three dimensional structure of object which is required in many high level vision applications. The focus measure presented has shown robustness in presence of noise as compared to the earlier focus measures. This new focus measure is based on an optical transfer function using Discrete Cosine Transform and its results are compared with the earlier focus measures including Sum of Modified Laplacian (SML) and Tenenbaum focus measures. With this new focus measure, the results without any noise are almost similar in nature to the earlier focus measures however drastic improvement is observed with respect to others in the presence of noise. The proposed focus measure is applied on a test image, on a sequence of 97 simulated cone images and on a sequence of 97 real cone images. The images were added with the Gaussian noise which arises due to factors such as electronic circuit noise and sensor noise due to poor illumination and/or high temperature. 6056-09, Session 3 Formation of stereoscopic image pairs from a sequence of frames M. A. Wessels, Dimensional Imaging, LLC Under certain conditions of relative motion between an object and an observer, stereoscopic image pairs may be constructed from a sequence of monoscopic frames depicting the object at specific points in time. Pure transverse relative motion will produce slightly displaced views from frame to frame in the sequence. These displacements are the source of the stereoscopic separation necessary for stereoscopic depiction.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
87
Conf. 6056: Three-Dimensional Image Capture and Applications VI
In general, the displacements occur around a stereoscopic axis that may lie at any orientation in the frame. The algorithm determines this axis, and then rotates the image pair to set the axis to lie in the vertical direction. Actual motion may include rotations of the object itself, which are superimposed upon the gross transverse motion. Non-contributing rotations are identified and removed. The starting point of this process is a displacement map which shows by how much specific points in a frame have shifted with respect to their counterparts in the corresponding reference frame. Pairs which exhibit good stereoscopic separation will yield displacement maps characterized by a strong gradient over the field. Analysis of the orientation of this gradient yields the stereoscopic axis. 6056-10, Session 3 3D model generation using unconstrained motion of a hand-held video camera C. Baker, C. H. Debrunner, M. Whitehorn, PercepTek, Inc. We have developed a shape and structure capture system which constructs accurate, realistic 3D models from video imagery taken with a single freely moving handheld camera, without the need for an expensive custom inertial measurement system. Using an inexpensive off the shelf hand-held video camera, we demonstrate the feasibility of fast and accurate generation of these 3D models at a very low cost. In our approach the operator films a scene while observing some very simple camera motion constraints. Our system identifies and tracks high interest image features and computes the relative pose of the camera using a RANSAC-based approach to solve for the camera pose and 3D structure which best describes the feature motion. Once we have the pose for many frames in the video sequence we perform correlation-based stereo to obtain dense point clouds. These point clouds are integrated into a 3D model, using an octree structure for efficient representation. The integration technique enables filtering based on voxel occupancy counts which eliminates many stereo outliers and results in an accurate, 3D model which may be inspected from novel viewpoints. This paper describes our approach in detail and shows reconstructed results of both indoor and outdoor scenes. 6056-11, Session 3 3D from arbitrary 2D video I. A. Ideses, L. P. Yaroslavsky, Tel-Aviv Univ. (Israel) In this paper we present a method to synthesize 3D video from arbitrary 2D video. The 2D video is analyzed by computing frameby-frame motion maps. For this computation several methods were tested, among them are optical flow, MPEG motion vectors extraction and correlation based target location. Using the computed motion maps, the video undergoes analysis and the frames are segmented to provide object-wise depth ordering. The frames are then used to synthesize stereo pairs. This is performed by resampling of the image on a grid that is governed by the depth-map. In order to improve the quality of the synthetic video and in order to enable 2D viewing where 3D visualization is not possible, several techniques for image enhancement are used. In our test case, anaglyph projection was selected as the 3D visualization method, since this method is mostly suited to standard displays. The drawback of this method is ghosting artifacts. In our implementation we minimize these unwanted artifacts by modifying the computed depth-maps using non-linear transformations. Defocusing was also used to counter such artifacts. Our results show that non-linear operations on the depth-maps enable synthesis of high quality 3D videos.
6056-12, Session 3 Nonintrusive viewpoint tracking for 3D for perception in smart video conference X. Desurmont, I. Ponte, J. Meessen, J. Delaigle, Multitel A.S.B.L. (Belgium) Globalisation of people's interaction in the industrial world and the ecological cost of transport make video-conference an interesting solution for collaborative work. However, the lack of immersive perception makes video-conference not appealing. TIFANIS teleimmersion system was conceived to let users interact as if they were physically together. In this paper, we focus on an important feature of the immersive system: the automatic tracking of the user's point of view in order to render correctly in his display the scene from the other site. Viewpoint information has to be computed in a very short time and the detection system should be no intrusive, otherwise it would become cumbersome for the user. The viewpoint detection system consists of several modules. First, an analysis module identifies and follows regions of interest (ROI) where faces are detected. We will show the cooperative approach between spatial detection and temporal tracking. Secondly, an eye detector finds the position of the eyes within faces. Then, the 3D positions of the eyes are deduced using stereoscopic images from a binocular camera. Finally, the 3D scene is rendered in real-time according to the new point of view. The performances of the system are evaluated in terms of accuracy of face and eyes 3D positions in various sequences. 6056-13, Session 3 Internal shape-deformation invariant 3D surface matching using 2D principal component analysis M. Celenk, I. Al-Jarrah, Ohio Univ. This paper describes a method that overcomes the problem of internal deformations in three-dimensional (3D) range image identification. Internal deformations can be caused by several factors including stereo camera-pair misalignment, surface irregularities, active vision methods' incompatibilities, image imperfections, and changes in illumination sources. Most 3D surface matching systems suffer from these changes and their performances are significantly degraded unless deformations' effect is compensated. Here, we propose an internal compensation method based on the two-dimensional (2D) principal component analysis (PCA). The depth map of a 3D range image is first thresholded using Otsu's optimal threshold selection criterion to discard the background information. The detected volumetric shape is normalized in the spatial plane and aligned with a reference coordinate system for rotation-, translation- and scaling-invariant classification. The preprocessed range image is then divided into 16x16 sub-blocks, each of which is smoothed to minimize the local variations. The 2DPCA is applied to the resultant range data and the corresponding principal vectors are used as the characteristic features of the object to determine its identity in the database of pre-recorded shapes. The system's performance is tested against the several 3D facial images possessing arbitrary deformation. Experiments have resulted in 92% recognition accuracy for the GavaDB 3D-face database entries and their Gaussian- or Poissontype noisy versions.
88
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6056: Three-Dimensional Image Capture and Applications VI
6056-14, Session 3 Digital Hammurabi: design and development of a 3D scanner for cuneiform tablets D. V. Hahn, D. Duncan, K. Baldwin, J. Cohen, B. Purnomo, Johns Hopkins Univ. Cuneiform is an ancient form of writing in which wooden reeds were used to impress shapes upon moist clay tablets. Upon drying, the tablets preserved the written script with remarkable accuracy and durability. There are many problems associated with studying these artifacts. For example, photographic records of the tablets many times prove to be inadequate as they lack the ability to alter the lighting conditions and view direction. As a solution, we describe a 3D scanner capable of acquiring the shape, color, and reflectance of a tablet as a complete 3D object. This data set could then be stored in an online library and manipulated by rendering software that would allow a user to obtain any view and lighting direction. The scanner utilizes a camera and telecentric lens to acquire images of the tablet under varying controlled illumination conditions. Image data are processed using photometric stereo and structured light techniques to determine the tablet shape; color information is reconstructed from primary color monochrome image data. The scanned surface is sampled at 26.8 µm lateral spacing and the height information is calculated on a much smaller scale. Scans of adjacent tablet sides are registered together to form a 3D surface model. 6056-15, Session 3 Three-dimensional surface reconstruction for evaluation of the abrasion effects on textile fabrics A. O. Mendes, P. T. Fiadeiro, R. A. Miguel, Univ. da Beira Interior (Portugal) The evaluation of the abrasion effects becomes important on the textile industry because abrasion is responsible for many surface changes that occur in garments. In particular, the pilling formation is a phenomenon caused by abrasion that affects fabrics more significantly altering their surface severely. The present work presents an optical method that enables topographic reconstructions of textile fabric samples and consequently, it is possible to evaluate and quantify the pilling formation resulted from their topographic changes. The proposed method is based on optical triangulation basically operating like a scanning system (digitiser) or a 3D scanner. A fabric sample is holded by a linear motorized stage and, when is on movement, a thin light stripe performs a full scan of the fabric sample surface. Simultaneously, images of the light stripe profile are being acquired and stored on a computer for further processing. The procedures for image data acquisition, storage and processing are carried out using specific algorithms written in the MatLab programming language. Finally, with the available processed data, it is possible a three-dimensional surface reconstruction and a quantification of the pilling formation of the analyzed fabric samples. 6056-16, Session 3 3D environment capture from monocular video and inertial data R. Clark, M. Lin, C. J. Taylor, Acuity Technology This paper presents 3D environment reconstruction from monocular video augmented with 6-DOF inertial data. One application targets sparsely furnished room interiors using high quality, normal field of view, handheld video, and linear accelerations and angular velocities from an attached inertial measurement unit. A second application
targets natural terrain with manmade structures, using heavily compressed, narrow field of view, aerial video, and position and orientation data from the aircraft navigation system. In both applications, the 6-DOF spatial offsets between the camera and inertial data are initially unknown, and only a small fraction of the scene is visible in any one video frame. We start by estimating sparse structure and motion from 2D feature tracks using a Kalman filter and/or bundle adjustment. The first application additionally incorporates a weak assumption of bounding perpendicular planes (estimated via RANSAC) to minimize a tendency to drift, while the second application requires tight integration of the navigational data to alleviate the poor conditioning caused by the narrow field of view. This is followed by dense structure recovery via graphcut-based multiview stereo. Finally, input images are texture-mapped onto the meshed 3D surface for rendering. We show sample experimental results from novel viewpoints. 6056-18, Session 3 The effects of different shape-based metrics on identification of military targets from 3D ladar data G. J. Meyer, J. Weber, Air Force Research Lab. The choice of shape metrics is important to effectively identify threedimensional targets. The performance (expressed as a probability of correct classification) of three metrics using point clouds of military targets rendered using Irma, a government tool that simulates the output of an active ladar system, is compared across multiple ranges, sampling densities, target types, and noise levels. After understanding the range of operating conditions a classifier is expected to see in the field, a process for determining the upperbound of a classifier and the significance of this result is assessed. Finally, the effect of sampling density and variance in the position estimates on classification performance will be examined to make intelligent system level design tradeoffs. 6056-19, Session 4 Digital 3D reconstruction of George Washington A. Razdan, Arizona State Univ. PRISM is a focal point of interdisciplinary research in geometric modeling, computer graphics and visualization at Arizona State University. Many projects in the last ten years have involved laser scanning, geometric modeling and feature extraction from such data as archaeological vessels, bones, human faces, etc. This talk will briefly outline the history and development of work at PRISM followed by a recently completed project on the 3D reconstruction of George Washington (GW). The project brought together forensic anthropologist, digital artists and computer scientists in the 3D digital reconstruction of GW at 57, 45 and 19 including detailed heads and bodies. Although many other scanning projects such as the Michelangelo project have successfully captured fine details via laser scanning, our project took it a step further, i.e. to predict what that individual (in the sculpture) might have looked like both in later and earlier years, specifically the process to account for reverse aging. Our base data was GW's face mask at Morgan Library and Hudon's bust of GW at Mt Vernon, both done when GW was 53. Additionally we scanned the statue at the Richmond's Capitol, various dentures, and other items. Other measurements came from clothing and even portraits of GW. The digital GWs were then milled in high density foam for a studio to complete the work. These will be unveiled at the opening of the new education center at Mt Vernon in fall 2006.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
89
Conf. 6056: Three-Dimensional Image Capture and Applications VI
6056-20, Session 4 The study of craniofacial growth patterns using 3D laser scanning and geometric morphometrics M. Friess, Anthrotech Inc. Throughout childhood, braincase and face grow at different rates and therefore exhibit variable proportions and positions relative to each other. Our understanding of the direction and magnitude of these growth patterns is crucial for many ergonomic applications and can be improved by advanced 3D morphometrics. The purpose of this study is to investigate this known growth allometry using 3D imaging techniques. The geometry of the head and face of 840 children, aged 2 to 19, was captured with a laser surface scanner and analyzed statistically. From each scan, 18 landmarks were extracted and registered using General Procrustes Analysis (GPA). GPA eliminates unwanted variation due to position, orientation and scale by applying a least-squares superimposition algorithm to individual landmark configurations. This approach provides the necessary normalization for the study of differences in size, shape, and their interaction (allometry). The results show that throughout adolescence, boys and girls follow a different growth trajectory, leading to marked differences not only in size, but also in shape. These differences can be observed during early childhood, but become most noticeable after the age of 13 years, when craniofacial growth in girls slows down significantly, whereas growth in boys continues for at least 3 more years. 6056-21, Session 4 A three-dimensional analysis of the geometry and curvature of the proximal tibial articular surface of hominoids E. K. Landis, P. A. Karnick, Arizona State Univ. This study uses new three-dimensional imaging techniques to compare the articular curvature of the proximal tibial articular surface in hominoids. It has been hypothesized that the curvature of the anteroposterior contour of the lateral condyle in particular can be used to differentiate humans and apes and reflect locomotor function. This study draws from a large comparative sample of extant hominoids to obtain quantitative curvature data. Threedimensional models of the proximal tibiae of 26 human, 15 chimpanzee, 15 gorilla, 17 orangutan, 16 gibbon and four Australopithecus fossil casts (AL 129-1b, AL 288-1aq, AL 333x-26, KNM-KP 29285A) were acquired with a Cyberware Model 15 laser digitizer. Curvature analysis was accomplished using a software program developed at Arizona State University's Partnership for Research In Stereo Modeling (PRISM) lab, which enables the user to extract curvature profiles and compute the difference between analogous curves from different specimens. Results indicate that the curvature of chimpanzee, gorilla and orangutan tibiae is significantly different from the curvature of human and gibbon tibiae, thus supporting the hypothesized dichotomy between humans and great apes. The non-significant difference between humans and gibbons is surprising given locomotor differences between these taxa. All four Australopithecus tibia were aligned with the great apes.
6056-22, Session 4 New approach in curve matching technique and its implications on human evolution research H. Vahdati, P. A. Karnick, Arizona State Univ. Our research focuses on midfacial prognathism (MFP) a welldocumented Neanderthal trait. Understanding the pattern of MFP should help to answer questions arising from the modern human origins controversy in that well-studied area. This research should help to determine whether or not continuity is apparent in the midfacial region of the European Middle and Upper Pleistocene hominids that bracket Neanderthals in time. This research produces geometric morphometric three-dimensional analyses of the surface data. The data are collected from the midface of modern humans and fossil humans using a computer scanner. An original software (designed for this purpose), which inputs laser-scanned data from the surface of the skulls, has been employed to transform the raw facial data to the comparable curves to investigate degrees of similarity and matching between the mid facial curves of different humans. Our initial study (samples including of modern humans and fossil humans) provides us with a not very significant divergent between Neanderthals and early modern humans in Europe. 6056-23, Session 4 Point cloud-based 3D head model classification using optimized EGI X. Tong, H. Wong, B. Ma, City Univ. of Hong Kong (Hong Kong China) In this paper, we apply a new approach into the classification of the 3D head models, which are represented in the format of point cloud, and get a satisfying result. In the experiment, we have the following classification rates: 71.3% for the original EGIs, 70.6% for 2D FDA and 81.7% for the 2D PCA. As a summary, the contributions of this paper mainly include the following two aspects: first, we extend the application of 2D subspace analysis to 3D object recognition by characterizing a 3D object model with a 2D image; second, for the 3D head model classification problem, by using 2D subspace analysis, we can achieve higher classification rate, and more importantly reduce the computation cost. The reduction in computation cost is embodied at two aspects: One is that optimized EGI has smaller dimension than the original EGI does which facilitates the subsequent classification, and the other is that only solving an eigen-system with a small dimension is required. 6056-24, Session 4 3D face structure extraction using shape matching morphing model F. Xue, X. Ding, Tsinghua Univ. (China) In general, the point correspondence and automatic face structure extraction are challenging problems. This is due to the fact that automatic extraction and matching of a set of significant feature points on different image views on the face, which are needed to recover the individual's 3-D face modal, is a very hard machine task. In this paper, in order to bypass this problem, our method recovers both the pose and the 3-D face coordinates using shape matching morphing model and iterative minimization of a metric based on the structure matching. A radial basis function (RBF) in 3-D is used to morph a generic face into the specific face structure and shape context (SC) is used to descript point shape. Basing on RBF and SC, shape distance is used to measure the similarity of two shapes. Experiment results are shown for images of real faces and promising result are obtained.
90
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6056: Three-Dimensional Image Capture and Applications VI
6056-25, Session 5 Posture and re-positioning considerations of a torso imaging system for assessing scoliosis P. O. Ajemba, N. G. Durdle, Univ. of Alberta (Canada); D. L. Hill, J. V. Raso, Glenrose Rehabilitation Hospital (Canada) The influence of posture and re-positioning (sway and breathing) on the accuracy of a torso imaging system for assessing scoliosis was evaluated. The system comprised of a rotating positioning platform and two laser digitizers. It required four partial-scans taken at 900 intervals over 10 seconds to generate four complete torso scans. Its accuracy was previously determined to be 1.1±0.9mm. Ten evenly spaced cross-sections obtained from forty scans of five volunteers in four postures (free-standing, holding side supports, holding front supports and with their hands on their shoulders) were used to assess the variability due to posture. Twenty cross-sections from twenty scans of two volunteers holding side supports were used to assess the variability due to positioning. The variability due to posture was less than 4mm at each cross-section for all volunteers. Variability due to sway ranged from 0-3.5mm while that due to breathing ranged from 0-3mm for both volunteers. Holding side supports was the best posture. Taking the four shots within 12 seconds was optimum. As major torso features that are indicative of scoliosis are larger than 4mm in size, the system could be used in obtaining complete torso images used in assessing and managing scoliosis. 6056-26, Session 5 Reverse engineering and rapid prototyping techniques to innovate prosthesis socket design G. Magrassi, G. Colombo, M. Bertetti, D. Bonacini, Politecnico di Milano (Italy) The paper investigates an innovative approach totally based on digital data to optimize lower limb socket prosthesis design. This approach uses a stump's detailed geometric model and provides to substitute the plaster cast obtained through the traditional manual methodology with a physical model, realized with Rapid Prototyping technologies; this physical model will be used for the socket lamination .We present a methodology for stump's digital geometric model 3D reconstruction able to describe with high accuracy and detail the complete structure subdivided into bones, soft tissues, muscular masses and dermis. Some different technologies are used for stump imaging and acquisition: non contact laser technique for external geometry acquisition, CT and MRI imaging technologies for the internal structure, the first one devoted to bones geometrical model, the last fit for soft tissues and muscles. We discuss about problems related to 3D geometric reconstruction: the relative position of patient and stump for the different acquisitions, markers' definition on the stump to identify landmarks, alignment's strategies for the different digital models, in order to define a protocol procedure answering for the requested accuracy for socket's realization. Some case-studies demonstrate the methodology and the obtained results. 6056-27, Session 5 4D data processing for dynamic human body analysis R. Sitnik, A. M. Filipczak, Politechnika Warszawska (Poland) It seems that the next generation of a full 3D optical scanning systems will be able to measure objects in motion. According to this it will be necessary to use other data processing and representation methods. We propose our own solution which is based on using an
arbitrary mesh which is scaled and wrapped around a merged point clouds obtained from the measurements instead of a standard point cloud representation. This method was invented for a project of motion analysis of human legs. The first step of our 4D data processing chain is scanning and merging points clouds obtained from four directions. After that we detect some selected areas on the object's surface and it's approximate pose/position. We create a simple bone system and using it we deform an arbitrary mesh to match the point cloud. The last step is wrapping this mesh around the point cloud. This is performed by calculating the best fitting surfaces to the selected points of the cloud and projecting each vertex of the mesh onto them. This kind of scanning data representation can have a very wide range of application - industry, medicine and multimedia. We present some details of our solution, exemplary results and finally some conclusions with a short discussion of our future work in this field. 6056-28, Session 5 Measuring human movement for biomechanical applications using markerless motion capture L. Mьndermann, S. Corazza, A. M. Chaudhari, T. P. Andriacchi, Stanford Univ.; A. Sundaresan, R. Chellappa, Univ. of Maryland/ College Park The development of methods for the capture of human movement is motivated by the desire to understand normal and pathological joint mechanics without the artifacts associated with standard markerbased motion capture techniques such as soft tissue artifacts and the risk of artificial stimulus of taped-on or strapped-on markers. In this study, the advancement of markerless human motion capture is described in the context of biomechanical applications. Several algorithms for accurate markerless motion capture are explored systematically in both a virtual and experimental environment, and results for real data are presented. The implementation of this new technology offers the promise for simple, time-efficient, and potentially more meaningful assessments of human movement in research and clinical practice. 6056-29, Poster Session Development of measurement system of three-dimensional shape and surface reflectance T. Miyasaka, K. Araki, Chukyo Univ. (Japan) We describe a three-dimensional measurement system which can acquire not only three-dimensional shapes of target objects but also these surface reflectance parameters. The system is constructed by one or some digital cameras, digital projector, and a computer which controls camera and projector. For 3-D geometrical reconstruction, we use well known gray code structured light method. The method projects gray code light patterns from the projector and obtain illuminated scenes by a camera. We add additional light patterns for surface reflectance measurement. These patterns are all white and gray light pattern. To recover complete shape of the target object, the object is measured from various viewpoints repeatedly, or measured repeatedly from fixed viewpoint while be moving by hand or turn table. To end the measurement, relative positions of each obtained range data are calculated by ICP algorithm. For each small region of the target object surface, we calculate reflectance parameters from surface normals, viewpoint (camera viewpoint), and light position (the projector viewpoint). Enough sampling of these three information sources are obtained for each small surface, we estimate reflectance parameters for each surface points.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
91
Conf. 6056: Three-Dimensional Image Capture and Applications VI
We demonstrate this geometrical and reflectance measurement method by experiments for fewer objects. 6056-30, Poster Session Use of laser 3D digitizer in data collection and 3D modeling of anatomical structures K. Tse, Univ. of New South Wales (Australia); H. Van Der Wall, Concord Repatriation General Hospital (Australia); D. H. Vu, Univ. of New South Wales (Australia) A laser digitizer (Konica-Minolta Vivid 910) is used to obtain 3dimensional surface scans of anatomical structures with a maximum resolution of 0.1mm. Placing the specimen on a turntable allows multiple scans all-around because the scanner only captures data from the portion facing its lens. A computer model is generated using 3D modeling software such as Geomagic. The 3D model can be manipulated on screen for repeated analysis of anatomical features, a useful capability when the specimens are rare or inaccessible (museum collection, fossils, imprints in rock formation...). As accurate measurements can be performed on the computer model, instead of taking measurements on actual specimens only at the archeological excavation site e.g., a variety of quantitative data can be later obtained on the computer model in the laboratory as new ideas come to mind. Our group had used a mechanical contact digitizer (Microscribe) for this purpose, but with the surface digitizer, we have been obtaining data sets more accurately and more quickly. 6056-31, Poster Session Volume intersection with imprecise camera parameters S. Sakamoto, K. Shoji, H. Iwase, F. Toyama, J. Miyamichi, Utsunomiya Univ. (Japan) Volume intersection is one of the simplest techniques for reconstructing 3D shapes from 2D silhouettes. 3D shapes can be reconstructed from multiple view images by back-projecting them from the corresponding viewpoints and intersecting the resulting solid cones. The camera position and orientation (extrinsic camera parameters) of each viewpoint with respect to the object are needed to accomplish reconstruction. However, even a little variation in the camera parameters makes the reconstructed 3D shape smaller than that with the exact parameters. The problem of optimizing camera parameters deals with determining exact ones from approximate ones and multiple silhouette images. This paper examines attempts to optimize camera parameters by reconstructing a 3D shape via the method of volume intersection. Reprojecting the reconstructed 3D shape to image planes, the camera parameters are determined by finding the projected silhouette images that result in minimal lost area when compared to the original silhouette images. For relatively large displacement of camera parameters we propose a method repeating the optimization using dilated silhouettes which gradually shrink to original ones. Results of simulation experiments using a VRML model show the effect of it. 6056-32, Poster Session Development of ultrathin three-dimensional image capturing system K. Yamada, H. Mitsui, T. Asano, Hiroshima Institute of Technology (Japan); H. Takahashi, Osaka City Univ. (Japan); J. Tanida, Osaka Univ. (Japan) We have developed the ultra thin three dimensional image capture system. The system uses a micro-lens array to form multiple images, which are captured on a photo-detector array. Digital processing of
the captured multiple images is used to extract the surface profile. Preliminary experiments were executed on an evaluation system to verify the principles of the system. The system consists of three components: a micro-lens array, a signal separator, and a photodetector array. Each micro-lens sends optical signals to the multiple photosensitive cells on the photo-detector array, which comprises an imaging unit. To prevent cross talk between adjacent lens units, an opaque wall is inserted as the signal separator. For the photodetector array, conventional CCD chip or a CMOS image sensor is used. An experimental system consists of a compound imaging system and a CCD image sensor. The diameter and the focal length of micro-lens are 500µm and 1.3mm, respectively. The high and thickness of the signal separator are 800µm and 20-30µm, respectively. In this paper, a compound-eye imaging system and post-processing are employed. Experimental results verify the principle of the proposed method and show the potential capability of the proposed system architecture. 6056-33, Poster Session Run-based volume intersection for shape recovery of objects from their silhouettes K. Shoji, S. Sakamoto, H. Iwase, F. Toyama, J. Miyamichi, Utsunomiya Univ. (Japan) Volume intersection (VI) is a successful technique for reconstructing 3D shapes from 2D images (silhouettes) of multiple views. It consists of intersecting the cones formed by back-projecting each silhouette. The 3D shapes reconstructed by VI are called visual hull (VH). In this paper we propose a fast method obtaining the VH. The method attempts to reduce the computational cost by using a run representation for 3D objects called "SPXY table" that is previously proposed by us. It makes cones by back-projecting the 2D silhouettes to the 3D space through the centers of the lens and intersects them keeping the run representation. To intersect the cones of multiple views keeping the run representation, we must align the direction of runs representing the cones. To align them we use the method of swapping two axes of a run-represented object at the time cost of O(n) where n is a number of runs, which is also previously proposed by us. The results of expe!riments using VRML objects such as human bodies show that the proposed method can reconstruct a 3D object in less than 0.17 s at the resolution of 220 x 220 x 220 voxels from a set of silhouettes of 8 viewpoints on a single CPU. 6056-34, Poster Session A prototype system for 3D measurement using flexible calibration method M. Fukuda, T. Miyasaka, K. Araki, Chukyo Univ. (Japan) We developed an easy-to-use calibration method that may calibrate a projector-camera-system with high accuracy at short time, and implemented the measurement system based on that calibration method were implemented. In our system, the camera and the projector are calibrated in advance by Zhang's calibration method. The measurement procedure in our system is as follows. The calibrated camera and the calibrated projector are put suitably in front of the calibration plane. And the relative pose between the camera and the projector is may compute by just projecting some light patterns (the horizontal and vertical gray-code pattern) from the projector onto the plane and taking those images by the camera. Then, this system measures by the gray-code pattern projection. Since the calibration of the system is easy, the layout of the system may change freely depending on the measurement target and the measurement location. In spite of that this system may be calibrated easily, this system may obtain range data in high accuracy of an error about 0.1%.
92
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6056: Three-Dimensional Image Capture and Applications VI
6056-35, Poster Session Estimation of object motion with known structure in moving camera H. J. Kwon, N. Hur, S. Lee, Electronics and Telecommunications Research Institute (South Korea) In this paper, we have proposed a method to estimate object's motion in a dynamic environment. Structure from Motion(SfM), the traditional method, has assumed that camera or object is fixed and extracts its motion. However, in case of both of them moving, SfM is not working in object area because a moving object prevents structure recovery from triangulation between two views. To solve this problem, traditional methods use stereo camera or calibrated camera. We focus on a possibility of uncalibrated camera setting using only minimum 3D information, relative scale of the scene and object. We have tested our methods with synthetic images and real images. To estimate object's 3D motion, we just used relative scale of background and object, and did not necessarily use other information such as metric scale of the scene and the object. In our simulation results for test images, proposed method has estimated object's motion successfully. Main contribution in the proposed method is the estimation of object's 3D motion and then the registration of moving objects and a moving camera in a single space, which is not clearly shown in traditional uncalibrated SfM and pose estimation. 6056-36, Poster Session Synthesizing wide-angle and arbitrary viewpoint images from a circular camera array N. Fukushima, T. Yendo, T. Fujii, M. Tanimoto, Nagoya Univ. (Japan) We propose a technique of Imaged-Based Rendering using a circular camera array. Recording a scene as surrounding, we can synthesize a more dynamic arbitrary viewpoint images and wideangle images like a panorama. This method is based on Ray-Space, one of the image-based rendering. Ray-Space is described by the position (x, y) and a direction (theta, phi) of the ray's parameter that passes a reference space. Thus, the user can change viewpoint freely. However, rigorous Ray-Space rule require so many images, thousands of images. In this research, we focus the case of lack of ray in Ray-Space for the problem of huge amount of data. Thus, rays interpolation technique becomes the key. We want to interpolate these rays along the source of ray in the 3D field. 3D modeling method solve this problem, however, moldering is too difficult to rendering photorealistic image. Thus, we propose the novel method of estimating the depth of ray in the virtual view. This technique search the source of ray using a periodic function on Ray-Space established by circular camera array. Finally, we can synthesize quality images with low cost like Image-Based Rendering. 6056-37, Poster Session 3D urban scene reconstruction from highresolution IKONOS stereo images M. Fedi, T. Riadh, B. Ziad, SUPCOM (Tunisia) High resolution IKONOS stereo imagery is very useful for 3D feature extraction. The IKONOS stereo product can provide a 1-meter resolution stereo pairs in epipolar geometry. The provided epipolar geometry simplifies considerably the matching step of the two
images allowing the end user to process directly the stereo pairs without additional re-sampling or rectification of the original images. However, having a very high-resolution rectified stereo image pair of an urban scene, the matching of the content of the scene remains still a difficult problem. Depth discontinuities, large occluded areas, mixture of textured and non-textured areas present in the urban scenes, are some of the problems to take care of. Many different approaches were overtaken to resolve these problems with all the constraints that can be brought with. We can classify these approaches in two large categories: area based and feature based stereo. Dealing with area based algorithms we can talk about cross-correlation based or region growing. While for feature based algorithms can deal with edge and/or corner detection and texture region based. Each of these approaches had its merits and disadvantages dependent on the nature of the matching problem. Many suggested combining the relevant approaches in a cooperative fashion. In this paper, we propose a cooperative method between many approaches to get the better of each one. The algorithm consists first on a pre-treatment of the stereo pair by applying an anisotropic diffusion on them. Then, we used non-redundant complex wavelet to get the coarse scales in order to perform the matching process in a hierarchical way. Gabor filters are then applied at each scale to get texture based segmentation. This segmentation is combined with an edge detection to extract the different existing urban objects in the images and features vectors that describe them. Finally and using their feature vectors, we performed the matching process of the obtained objects from the coarsest scale to the finest one to get the disparity card. Results show that, in the case of 3D urban scenes reconstruction, combining texture features with edge features is well suited for matching the stereo images. 6056-38, Poster Session Procedure and algorithm of 3D reconstruction of large-scale ancient architecture S. Xia, Y. Zhu, X. Li, Wuhan Univ. (China) Techniques of 3D reconstruction and photogrammetry provide an efficient way for documentation and protection of ancient architecture. Close-range photogrammetry can be used to survey geometric figure of 3D model accurately. Photos of 3D object (such as granite) are taken with a metric camera, which has known and adjusted elements of interior orientation. Sufficient feature points on 3D object are marked to represent the profile of the solid. Then DLT (Direct Linear Transform) sets up direct linear relation between the photo coordinates and object space coordinates. Space coordinates of the object are input into CAD to restore the profile. At this stage, a new algorithm for generating spatial convex polyhedron is presented and realized in our research. 3D TIN (Triangulated Irregular Network) of the object can be obtained efficiently. Model connection and optimization are performed to reduce duplicate storage of points and edges of the model. Realistic material can be attached to the 3D model in 3DMAX. At last rendering and animation of the 3D model are completed and we got the reconstructive model. After assembly of 3D component model of the architecture, we can realize the 3D reconstruction of large-scale ancient architecture successfully.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
93
Conf. 6056: Three-Dimensional Image Capture and Applications VI 6056-41, Poster Session Real-time 3D image-guided patient positioning in radiation therapy D. Liu, Henry Ford Health System; G. Yin, Genex Technologies, Inc. (China); S. Li, Henry Ford Health System Patient positioning in modern radiotherapy is becoming far more important because a small positioning error may result in missing target and irradiating normal tissues in treatment of small lesions. Clinical outcome of radiotherapy can potentially be improved by increasing the precision of tumor localization and dose delivery during the treatment. In this paper an accurate and precise patient positioning system has been achieved through alignment of realtime three dimensional (3D) surface images with a reference surface image. The real-time 3D surface is captured using a state-of-art 3D stereovision system, and then is matched with the pre-defined reference image generated from treatment planning data. Positioning parameters are calculated by automatically aligning the real-time surface and the reference surface via a modified Iterative Closest Points (ICP) algorithm. Results from phantom experiments and clinical applications demonstrated the excellent efficacy of <2 minutes and the desired accuracy and precision of <1 mm in isocenter shifts and of <1 degree in rotations.
94
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6057: Human Vision and Electronic Imaging XI
Monday-Thursday 16-19 January 2006 Part of Proceedings of SPIE Vol. 6057 Human Vision and Electronic Imaging XI
6057-01, Session 2 Local luminance effect on spatial summation in the foveal vision and its implication on image artifact classification C. Chen, S. Y. Lin, H. G. Han, National Taiwan Univ. (Taiwan) We investigated the spatial summation effect on pedestals with difference luminance. The targets were luminance modulation defined by Gaussian functions. The size of the Gaussian spot was determined by the scale parameter ("standard deviation", ) which ranged from 0.075° to 0.6°. The local luminance pedestal (2° radius) had mean luminance ranged from 2.9 to 29cd/m2. The no-pedestal condition had a mean luminance 58cd/m2. We used a QUEST adaptive threshold seeking procedure and 2AFC paradigm to measure the target contrast threshold at different target sizes (spatial summation curve) and pedestal luminance. The target threshold decreased as the target area increased with a slope -1 on log-log coordinates. However, if the target size was large enough ( \>0.15°, there was little, if any, threshold reduction as the target size further increased. The spatial summation curve had the same shape at all pedestal luminance levels. The effect of the pedestal was to shift the summation curve vertically on log-log coordinates. Hence, the size and the luminance effects on target detection are separable. The visibility of the Gaussian spot can be modeled by a function with a form f(L)*g() where f(L) is a function of local luminance and g() is a function of size. 6057-02, Session 2 Evaluating contrast sensitivity S. Kitagura, L. W. MacDonald, London College of Communication (United Kingdom) The problem for proper rendering of spatial frequencies in digital imaging applications is to establish the relative contrast sensitivity of observers at suprathreshold contrast levels in typical viewing environments. In an experimental study two methods of evaluating spatial contrast sensitivity were investigated, using targets of graded tonal modulation, at which observers were asked to point to the perceived threshold locations. The results produced by these two methods were rather different from those of the classical methods of vision science, showing a much lower sensitivity over a broader range of spatial frequencies. These may be regarded as complementary to CSF data derived from single-frequency Gabor stimuli and may prove to be better suited to the needs of practical imaging applications. 6057-03, Session 2 Spatio-velocity CSF as a function of retinal velocity using unstabilized stimuli J. L. Laird, M. R. Rosen, J. B. Pelz, E. D. Montag, Rochester Institute of Technology; S. J. Daly, Sharp Labs. of America, Inc. LCD televisions have LC response times and hold-type data cycles that contribute to the appearance of blur when objects are in motion on the screen. New algorithms based on studies of the human visual system's sensitivity to motion are being developed to compensate for these artifacts. This paper describes a series of experiments that incorporate eye-tracking in the psychophysical determination of spatio-velocity contrast sensitivity in order to build on the 2D spatiovelocity contrast sensitivity function (CSF) model first described by
Kelly and later refined by Daly. We explore whether the velocity of the eye has an effect on sensitivity and whether the model can be used to predict sensitivity to more complex stimuli. There were a total of five experiments performed in this research. The first four experiments utilized Gabor patterns with three different spatial and temporal frequencies and were used to investigate and/or populate the 2D spatio-velocity CSF. The fifth experiment utilized a disembodied edge and was used to validate the model. All experiments used a two interval forced choice (2IFC) method of constant stimuli guided by a QUEST routine to determine thresholds. The results showed that sensitivity to motion was determined by the retinal velocity produced by the Gabor patterns regardless of the type of motion of the eye. Based on the results of these experiments the parameters for the spatio-velocity CSF model were optimized to our experimental conditions. 6057-04, Session 2 A basis for cones B. V. Funt, W. Xiong, Simon Fraser Univ. (Canada) Why do the human cones have the spectral sensitivities they do? We hypothesize that the cones may have evolved to their present form because their sensitivities are optimal in terms of the amount of information they capture about the spectrum of incident light. As evidence in favor of this hypothesis, we compare the accuracy with which the incoming spectrum can be approximated by a threedimensional linear model based on the cone responses and compare this to the optimal approximations defined by models based on principal components analysis, independent component analysis, non-negative matrix factorization and non-negative independent component analysis. We introduce a new method of reconstructing spectra from the cone responses and show that the cones are almost as good as these optimal methods in estimating the spectrum. 6057-05, Session 2 Computational model of lightness perception in HDR imaging G. Krawczyk, K. Myszkowski, H. Seidel, Max-Planck-Institut fьr Informatik (Germany) An anchoring theory of lightness perception by Gilchrist et al. [1999] explains many characteristics of human visual system such as lightness constancy and its spectacular failures which are important in the perception of images. The principal concept of this theory is the perception of complex scenes in terms of groups of consistent areas (frameworks). Such areas, following the gestalt theorists, are defined by the regions of common illumination. The key aspect of the image perception is the estimation of lightness within each framework through the anchoring to the luminance perceived as white, followed by the computation of the global lightness. In this paper we provide a computational model for automatic decomposition of HDR images into frameworks. We derive a tone mapping operator which predicts lightness perception of the real world scenes and aims at its accurate reproduction on low dynamic range displays. Furthermore, such a decomposition into frameworks opens new grounds for local image analysis in view of human perception.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
95
Conf. 6057: Human Vision and Electronic Imaging XI
6057-47, Session 2 High-dynamic range scene compression in humans J. J. McCann, McCann Imaging Single pixel compression alters a particular input value to a unique output value, a look-up table. It is used in chemical and most digital photographic systems, having an S-shaped transforms to render high-range scenes onto low-range media. Post- receptor neural processing is spatial. (Dowling, Barlow, Kuffler, Hubel & Wiesel). Human vision does not render particular receptor-quanta catch as a unique response. Instead, because of spatial processing, the response to a particular quanta catch can be any color. Visual response is scene dependent. Stockham proposed an approach to model human-range compression using low-spatial frequency filters. Campbell, Ginsberg, Wilson, Watson, Daly and many others have developed independent spatial-frequency channel models. This paper describes experiments measuring suitable spatial-frequency filters for a variety of scenes. Given the radiances of each pixel in the scene and the observed appearances of objects in the image, one can solve for the best filters for that individual image. Low-dynamic range images with many white areas need no spatial filtering. Highdynamic-range images with many blacks or deep shadows require strong spatial filtering. Sun on the right and shade on the left requires directional filters. These experiments show that variable scene-dependent filters are necessary to mimic human vision. Although multiple-channel spatial-frequency filters can model human appearances, the problem still remains that an analysis of the scene is still needed to calculate the scene-dependent strengths of each of the channel filters. 6057-06, Session 3 A new metrics for definition of gaze area from the geometrical structures of picture composition M. Yamazaki, M. Kameda, Iwate Prefectural Univ. (Japan) The gaze area is changed by timeline when human observes picture. A change of this gaze area is defined as dynamic gaze flow. The dynamic gaze flow is used picture cording, image evaluation and computer vision effectively. Our purpose is to analyze the dynamic gaze flow by image only. We propose a new analysis method for the analysis of the gaze area which is analyzed the human visual characteristics to apply image processing. By the way, the technique with composition is the photograph and fine arts. Composition has the function which attract the human attention in some specific area. It is supposed that the dynamic gaze flow has a relationship with composition in picture. Our proposed method analyzes the composition which is based on the brightness of still picture as the primary factor of gazing, is evaluated by comparing the dynamic gaze flow of human which is measured with eye-mark recorder. As a result, it is clarified using several pictures that a course of the dynamic gaze flow is obtained by the analysis of picture composition.
6057-43, Session 3 Target salience and visual search on novel and familiar backgrounds K. McDermott, University of Nevada Reno; J. Mulligan, NASA Ames Research Center; G. Bebis, M. Webster, University of Nevada Reno The visual salience of a target depends on how the target differs from its background. We examined whether salience and the properties of visual search also depend on whether the backgrounds are novel or familiar. Eye movements were monitored while observers searched for a circular or elliptical element placed at random locations within a 30 by 40 deg background composed of a dense array of ellipses. The colors in the background varied randomly along either the LM or S cardinal axes, and also varied in luminance. On different trials the contrast of the target color relative to the background ranged from low (near or within the background color distribution) to high (far removed from the background color axis). Observers searched for the targets on a given background after adapting for 2 min to a sequence of LM or S varying backgrounds. Fixations during search were recorded with a CRS video eyetracker. On each background search times decreased with increasing target to background contrast, but were consistently faster when searching on backgrounds defined by the adapted color axis than by the orthogonal color axis. These results are consistent with an adaptation effect that increases the salience of novel colors by reducing the effective contrast of the background axis, and we analyze the distribution of eye movements to assess whether there are, in addition, changes in the search patterns on novel and familiar backgrounds. 6057-46, Session 3 Larry Stark and scan path S. R. Ellis, NASA Ames Research Ctr. No abstract available 6057-49, Session 3 Guiding the mind's eye: improving communication and vision by external control of the scanpath E. Barth, M. Dorr, M. Bцhme, Univ. zu Lьbeck (Germany); K. Gegenfurtner, Justus-Liebig-Univ. Giessen (Germany); T. Martinetz, Univ. zu Lьbeck (Germany) Larry Stark has emphasized that what we visually perceive is very much determined by the scanpath, i.e. the pattern of eye movements. Inspired by his view, we have studied the implications of the scanpath for visual communication and came up with the idea to not only sense the gaze but also guide it by using a special kind of gaze-contingent information display. Our goal is to integrate gaze into visual communication systems by measuring and guiding eye movements. We currently work with high-resolution natural videos. For guidance, we first predict a set of about 10 salient locations. We then change the probabilities for one of these candidates to be attended: for one candidate the probability is increased, for the others it is decreased. To increase saliency, for example, we add red dots that are displayed very briefly such that they are hardly perceived consciously. We experiment with further possibilities such as local zoom or jiggle. To decrease the probability we locally reduce the spatial and/or temporal frequency content. Again, if performed in a gaze-contingent fashion with low latencies, these manipulations of the video remain unnoticed. Overall, the goal is to find that real-time transformation of the video, which minimizes the difference between the actual and the desired scanpath. Applications are in the area of vision-based communication (better
96
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6057: Human Vision and Electronic Imaging XI
control of what information is conveyed) and augmented vision (guide a person's gaze by the gaze of an expert or a computer-vision system). We believe that our research is very much in the spirit of Larry Stark's views on visual perception and the close link between vision research and engineering. The research is based on collaboration between the Universities of Lьbeck and GieЯen and Sensomotoric Instruments GmbH in Teltow/ Berlin, and is supported by the German Federal Ministry of Education and Research as part of the interdisciplinary project ModKog - for details see http://www.inb.uni-luebeck.de/Itap/ 6057-07, Session 4 Effects of spatial correlations and global precedence on the visual fidelity of distorted images D. M. Chandler, K. H. S. Lim, S. S. Hemami, Cornell Univ. This study presents the results of a subjective rating experiment designed to investigate the effects of spatial correlations and the disruption of global precedence on the visual fidelity of distorted images. Using a psychophysical scaling paradigm, subjects placed distorted images along a linear, one-dimensional "distortion axis" in which physical distance corresponded to perceived distortion. The images were distorted at fixed levels of total distortion contrast via JPEG and JPEG-2000 compression, via Gaussian blurring, via additive white noise, and via direct quantization of the DWT subbands in a manner specifically designed to disrupt the global-tolocal integration of structure across scale-space; these methods were chosen to provide a reasonable representation of commonly encountered distortions. Results revealed that, at highly suprathreshold distortions contrasts, distortions which were spatially correlated with the images gave rise to lower visual fidelity as compared to additive white noise. Furthermore, among the types of distortions tested, those which disrupted the global precedence effect generally gave rise to the lowest visual fidelity. Based on these data, an algorithm is presented which estimates visual fidelity by using a measure of the degree to which the distortions are suprathreshold, and a measure of the degree to which the distortions disrupt global precedence. 6057-08, Session 4 Pseudo no reference image quality metric using perceptual data hiding A. Ninassi, P. Le Callet, F. Autrusseau, Univ. de Nantes (France) Regarding the important constraints due to subjective quality assessment, objective image quality assessment has recently been extensively studied. Such metrics are usually of three kinds, they might be Full Reference (FR), Reduced Reference (RR) or No Reference (NR) metrics. We focus here on a new technique, which recently appeared in quality assessment context: data-hiding-based image quality metric. Regarding the amount of data to be transmitted for quality assessment purpose, watermarking based techniques are considered as pseudo no-reference metric: A little overhead due to the embedded watermark is added to the image. Unlike most existing techniques, the proposed embedding method exploits an advanced perceptual model in order to optimize both the data embedding and extraction. A perceptually weighted watermark is embedded into the host image, and an evaluation of this watermark leads to assess the host image's quality. In such context, the watermark robustness is crucial; it must be sufficiently robust to be detected after very strong distortions. The watermark distortion must be proportional to the image's distortion. Our work is compared to existing standard RR and NR metrics in terms of both the correlation with subjective assessment and of data overhead induced by the mark.
6057-09, Session 4 Attention-based color correction F. W. M. Stentiford, Univ. College London (United Kingdom) The color of illumination of a scene can have a dramatic affect on the performance of image retrieval systems. In addition different imaging devices will produce widely different responses. As there is normally no control over the camera characteristics, image preprocessing, the brightness or the color of the illumination, and the surface reflectances, this becomes a serious problem for object recognition and will lead to apparently identical images being assessed as different by the machine. This paper proposes a new algorithm that extracts color correction parameters from pairs of images and enables the perceived illumination of one image to be imposed on the other. The algorithm does not rely upon prior assumptions regarding illumination constancy and operates between images that can be significantly different in content. The work derives from related research on visual attention and similarity in which the performance distributions of large numbers of randomly generated features reveal characteristics of images being analysed. Results are reported using structurally identical images as well as cropped images and those with different content. 6057-10, Session 4 Contrast enhancement of medical images using multiscale decomposition M. A. Trifas, J. M. Tyler, O. S. Pianykh, Louisiana State Univ. This paper presents the enhancement of images with multiscale methods. The methods are based on the Laplacian Pyramid and 2D wavelets. The basic approach in multiscale enhancement is to decompose the image into components that represent individual details, and to improve the contrast by operating on the details rather than on the whole original image. The human visual system is sensitive to the different spatial frequencies in an image. In particular, the plots for human visual frequency indicate that some frequencies effect visualization more than others and some frequencies are not important at all. Removing certain frequencies can even help emphasize those that remain (keeping the total image "energy" constant), and improve the quality of the image. We have studied the effects of the most common artifacts (such as blurring and noise) on the frequency content of each image. We have compared (using statistical parameters) the multiscale decompositions corresponding to a blurred or sharp medical image with that corresponding to an "ideal" image. Based on these comparisons, we have computed values of coefficients in different frequency bands that when applied to the components of the blurred or extremely sharp images, enhances the visualization of these images. 6057-11, Session 4 Human visual alpha stable models for digital halftoning A. J. Gonzбlez, J. Bacca Rodrнguez, G. R. Arce, Univ. of Delaware; D. L. Lau, Univ. of Kentucky Human visual system (HVS) modeling has become a critical component in the design of digital halftoning algorithms. Methods which exploit the characteristics of the HVS include the direct binary search (DBS) and optimized tone-dependent halftoning approaches. Several HVS models have been proposed in the literature, among them, the broadly used Nдsдnen's exponential model. As shown experimentally by Kim and Allebach, Nдsдnen's model is constrained in shape and richer models are needed in order to attain
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
97
Conf. 6057: Human Vision and Electronic Imaging XI
better halftone attributes and to control the appearance of undesired patterns. As an alternative, they proposed a class of HVS models based on mixtures of bivariate Gaussian functions. In this work, alpha stable functions, an elegant class of models richer than mixed Gaussians, are exploited. These are more efficient than Gaussian mixtures as they use less parameters to characterize the tails and bandwidth of the model. It is shown that a decrease in the model's bandwidth leads to homogeneous halftone patterns and conversely, models with heavier tails yield smoother textures. A frequency domain analysis of the halftones will show that this observation agrees with the blue noise model. These characteristics, added to their simplicity, make alpha stable models a powerful tool for HVS characterization. 6057-12, Session 4 Study of asthenopia caused by the viewing of stereoscopic images H. Hagura, Tokyo Institute of Technology (Japan) EI105 Human Vision and Electronic Imaging XI Recently the various glassless 3D imaging technologies for 3D TV, PC, PDA(Personal Digital Assistant) and cellular phone have been developed. The 3D images in these devices are often watched for long time. However, most of the observers who watch 3D images for long time are suffered from asthenopia, eye fatigue, or eyestrain. This study trys to find the fundamental reasons of the symptom by using MEG (magnetoencephalograpy) and other apparatus like a 3D optometer. Until now, many types of three dimensional (3D) imaging devices have been done. Especially in these days without troublesome by utilizing special 3D glasses the promising stereoscopic displays we can watch with naked eyes (glassless) have been widely used. Even if using such glassless 3D displays, there are not a few people who feel fatigue in their eyes by observing the 3D images for long time. The several studies regarding such asthenopia (eye fatigue) have been already done, and the verification experiments by using various ophthalmological devices have been conducted. However, there are so far no definitive theories or reasons for this kind of fatigue which have been clearly described. Generally speaking, it is said that it is due to the unnatural viewing with the different positions between accommodation and convergence of the eyes is needed on watching 3D images, in comparison with usual observation of 2D images. There are many cases that the person who has not been accustomed to such viewing feels fatigue. The researches have been mainly subjective and qualitative until now. Quantitative analyses in strict sense were not made yet. In this study, in order to verify what has happen inside eyes and brain on observing 3D images and how substantial asthenopia is occurred, the highest performance MEG with 440 channels (brain magnetic measurement equipment) which was newly developed has been used. Depending upon observation of 3D Images, it is hard to avoid the fact that asthenopia occurs in some extend, but in order to lighten, it is a necessity to show some kind of guideline for 3D image production. As for one of the last goal of this study, the guideline will be proposed by presenting the objective results measured as the foundation.
6057-13, Session 5 Perceptual image quality improvement for large screen displays F. Lebowsky, Y. Huang, H. Wang, STMicroelectronics (France) As large-scale direct view TV screens such as LCD flat panels or plasma displays become more and more affordable, consumers not only expect to buy a `big screen' but to also get `great picture quality' especially when the screen resolution reaches or exceeds the well known HDTV standard. But for quite some time we will still face a limited bandwidth for video over IP networks for example. Therefore, many images will be available at low resolution only. In addition, due to the size of the screen, its viewing distance is often reduced. To display the low resolution images on a large screen, it is desirable to re-scale to the screen resolution. Consequently, more artifacts related to digital signal processing, such as blurriness or jagginess, easily show above the threshold of visual detection. Since the human visual system shows a non-linear behavior towards the detection of these artifacts we elaborate some simple image operations that better adapt to the image quality constraints of the human visual system. As a result, low resolution images look more pleasing on large screen displays in terms of sharpness and show less ringing than common traditional approaches such as cubic spline interpolation or bi-cubic interpolation filters. 6057-14, Session 5 LCD motion-blur analysis, perception, and reduction using synchronized backlight flashing X. Feng, Sharp Labs. of America, Inc. One of the image quality issues of LC TV is the motion blur caused by (1) LCD slow temporal response, and, (2) non-impulse, "hold" characteristics of the display. In this paper, the LCD motion blur was mathematically modeled using a frequency domain analysis, where the motion of an object causes temporal component in the spatial/ temporal spectrum. The combination of eye tracking and the display temporal low pass filtering causes the perception of the motion blur. One way to reduce the motion blur is backlight flashing, where the shorter "on" duration reduces the display temporal aperture function, thus improves the temporal transfer function of the display. We implemented backlight flashing on a LCD with a backlight system consisting of an array of light emitting diodes (LED). The LED can be flashed on for a short duration after LCD reaches the target level. The effect of motion blur reduction was evaluated both objectively and subjectively. The objective experiment involves with a model of eye tracking and a sequence captured images using a high speed camera. The subjective study compares the motion blur to an edge with a simulated edge blur. The comparison of objective and subjective experiments shows a good agreement. 6057-15, Session 5 Human vision-based algorithm to hide defective pixels in LCDs T. R. Kimpe, S. Coulier, Barco N.V. (Belgium) Producing displays without pixel defects or repairing defective pixels is technically not possible at this moment. This paper presents a new approach to solve this problem: defects are made invisible for the user by using image processing algorithms based on characteristics of the human eye. The performance of this new algorithm has been evaluated using two different methods. First of all the theoretical response of the human eye was analyzed on a series of images and this before and after applying the defective pixel compensation algorithm. These results show that indeed it is possible to mask a defective pixel. A
98
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6057: Human Vision and Electronic Imaging XI
second method was to perform a psychovisual test where users were asked whether or not a defective pixel could be perceived. The results of these user tests also confirm the value of the new algorithm. Our "defective pixel correction" algorithm can be implemented very efficiently and cost-effectively as pixel-data-processing algorithms inside the display in for instance an FPGA, a DSP or a microprocessor. The described techniques are also valid for both monochrome and color displays ranging from high-quality medical displays to consumer LCD-TV applications. 6057-16, Session 5 Using optimal rendering to visually mask defective subpixels D. S. Messing, L. J. Kerofsky, Sharp Labs. of America, Inc. Visually Optimal Rendering is a subpixel-based method of rendering imagery on a colour matrix display that jointly maximises displayed resolution and minimises attendant colour aliasing. This paper first outlines the Constrained Optimisation framework we have developed for the design of Optimal Rendering Filters. This framework takes into account the subpixel geometry and colour primaries of the panel, and the luminance and chrominance Contrast Sensitivity Functions of the visual system. The resulting Optimal Rendering Filter Array can be designed for any regular 2D lattice of subpixels, including multi-primary lattices. The mathematical machinery of Visually Optimal Rendering is then applied to the problem of visually masking defective subpixels on the colour matrix display. The rendering filters that result are able to reduce single subpixel black defects for any primary colour to the point of near invisibility at normal viewing distances. These filters have the interesting property of being spatially varying. This property allows the Optimal Filter Array to intelligently modify the values surrounding a defect in a way that takes advantage of the visual system's different sensitivities to luminance and chrominance detail in order to best mask a defect. 6057-17, Session 6 Perceptual study of the impact of varying frame rate on motion imagery interpretability quality C. P. Fenimore, National Institute of Standards and Technology; J. M. Irvine, Science Applications International Corp.; D. Cannon, National Geospatial-Intelligence Agency; J. W. Roberts, I. Aviles, National Institute of Standards and Technology; S. A. Israel, Science Applications International Corp.; M. Brennan, The Boeing Co.; L. Simon, J. R. Miller, D. S. Haverkamp, Science Applications International Corp.; P. F. Tighe, M. Gross, Booz Allen Hamilton Inc. As part of development of a quality scale for motion imagery interpretability, this perceptual study measures frame rate effects for simple interpretation tasks. Using synthetic imagery permits full control of the contrast and speed of moving objects, motion complexity, the number of confusers, and the noise structure. To explore the detectibility threshold, contrast between the moving objects and the background is set at 5%, 2%, and 1%. Nine viewers detect or identify a moving synthetic "bug" in each of 288 10second clip. Frame rate, contrast, and confusers have a significant effect on image interpretibility, while the speed and background show no significant effect. Generally, there is a significant loss in correct detection and identification for frame rates below (but not above) 10 F/s. Increasing contrast improves detection and at high contrast, confusers do not affect detection. Confusers reduce detection of higher speed objects. Higher speed improves detection, but complicates identification, although this effect was small. Higher
speed makes detection harder at 1 Frame/s, but improves detection at 30 F/s. The low loss of quality at moderately lower frame rates may have implications for bandwidth limited systems. 6057-18, Session 6 Color preference and perceived color naturalness of digital videos C. C. Koh, J. M. Foley, S. K. Mitra, Univ. of California/Santa Barbara Five second clips of color chroma scaled digital videos were presented to naпve subjects. The subjects rated either the level of preference or naturalness produced by the color manipulation. Color chroma was manipulated in the CIE 1976 (L*u*v*)-space; no other color attributes were modified. The objectives were: (1) to determine how the scaling of color chroma affected color preference and how it affected the judged naturalness of colors, and (2) to determine what relationship exists between preference and naturalness. The mean opinion score (MOS) functions for preference (PMOS) and naturalness (NMOS) increased to a maximum and decreased as the mean chroma (MC) of the videos was increased. Both PMOS and NMOS peaked at relatively high MC levels, with NMOS peaking at the same or lower MC than PMOS for all video contents. The PMOS and NMOS functions for individual videos were approximated relatively well by simple Gaussians. However, the functions for different videos were displaced relative to one another on the chroma axis and the displacement between the naturalness and preference functions also varied with the video contents. 6057-19, Session 6 Stabilizing viewing distances in subjective assessments of mobile video M. D. Brotherton, British Telecommunications plc (United Kingdom); K. Brunnstrцm, Acreo AB (Sweden); D. Hands, British Telecommunications plc (United Kingdom) Improvements in audio and video compression techniques, handset display capabilities allied to improved network transmission speeds have resulted in the emergence of mobile video services. These new video services will have to meet the expectations of customers in terms of pricing, content and reproduction quality. The perceptual quality of mobile video is affected by a number of factors, including network performance, image size, frame rate, handset specification coding scheme and bit-rate. Given the options associated with generating and transmitting mobile video, there is an industry requirement for video quality measurement tools. These tools will be used to determine the best means of preparing and delivering video for quality-critical mobile services. To provide this measurement capability, there is activity aimed at developing objective perceptual quality measurement models. The Video Quality Experts Group (VQEG) is presently working on defining a series of multimedia subjective quality tests. The results from these subjective tests will be used to evaluate the performance of alternative multimedia perceptual quality methods, including methods for mobile video. This paper describes a subjective study aimed at identifying differences between stabilising or not stabilising viewing distances in subjective quality tests of low resolution video. A series of subjective tests conducted over two laboratories, employing either fixed viewing distances with a chin rest or not, were investigated. The focus of this work is to determine whether fixing viewing distances in subjective quality tests of mobile video will provide more reliable subjective data, especially when considering the aggregation of results between laboratories.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
99
Conf. 6057: Human Vision and Electronic Imaging XI
6057-20, Session 6 Predicting subjective video quality from separated spatial and temporal assessment R. R. Pastrana-Vidal, J. Gicquel, J. Blin, France Telecom R&D (France); H. Cherifi, Univ. de Bourgogne (France) During real time video communications over packet networks, various degradations can occur on spatial or temporal signal axes. The end-user may perceive loss of image clearness-sharpness and fluidity impairments on visual information. The overall perceived degradation may indeed be seen as a combined contribution of both perceptual axes. A significant perceptual interaction between spatial and temporal quality has been highlighted by a set of subjective quality assessment tests. We show that, at least in our experimental conditions, the overall visual quality can be estimated from independent spatial (clearness-sharpness) and temporal (fluidity) quality assessments. Four visual quality prediction models are presented. The models' predictions show significant correlation with mean opinion scores from a group of observers. The model showing the highest performance takes into account non linear human assessment characteristics. Our results lead to a better understanding of spatiotemporal interactions and they could be useful for the conception of automatic video quality metric. 6057-21, Session 6 Handling of annoying variations of performances in video algorithm optimization M. M. Nicolas, STMicroelectronics (France) Evaluation and optimization, with an ever increasing variety of material, are getting more and more time-consuming tasks in video algorithm development. An additional difficulty in moving video is that frame-by-frame perceived performance can significantly differ from real-time per¬ceived performance. This paper proposes a way to handle this difficulty in a more systematic and objective way than with usual long tuning procedures. We take the example of interpolation algo¬rithms where variations of sharpness or contrast look annoying in real-time whereas the frame- by- frame performance looks well acceptable. These variations are analyzed to get an objective measure for the real- time annoyance. We show that the reason for the problem is that most inter¬polation algorithms are optimized across intraframe criteria ignoring that the achievable intrinsic performance may vary from frame to frame. Our method is thus based on interframe optimization taking into account the measured annoyance. The optimization criteria are steered frame by frame depending on the achievable performance of the current interpolation and the achieved perfor¬mance in previous frames. Our policy can be described as "better be good all time than very good from time to time". The advantage is that it is automatically controlled by the compromise wished in the given application. 6057-22, Session 6 Structural similarity quality metrics in a coding context: exploring the space of realistic distortions A. Brooks, T. N. Pappas, Northwestern Univ. Perceptual image quality metrics have explicitly accounted for perceptual characteristics of the human visual system (HVS) by modeling sensitivity to subband noise as thresholds above which distortion is just-noticeable. Another class of quality metrics, known as Structural SIMilarity (SSIM), account for perception more implicitly with the assumption that the HVS is adapted for extracting
structural information (relative spatial covariance) from images. We evaluate the effectiveness of the Complex Wavelet SSIM (CWSSIM), a translation-insensitive SSIM extension, in the context of realistic distortions that arise from compression and error concealment in video transmission applications. In order to better explore the space of distortions, we propose models for typical distortions encountered in video compression/transmission applications. We also derive a multi-scale CWSSIM that effectively handles local mean shift distortions. Finally, we use CWSSIM to evaluate traditional DCT image compression (JPEG) versus modern waveletbased algorithms (SPIHT, JPEG2000) and find that CWSSIM generally agrees with perceptual intuition. 6057-23, Session 6 Lossy compression of high dynamic range images and video R. Mantiuk, K. Myszkowski, H. Seidel, Max-Planck-Institut fьr Informatik (Germany) Most common image and video formats have been designed to work with existing output devices, like LCD or CRT monitors. As the display technology makes a huge progress, these formats can no longer represent the data that the new devices can display. Therefore a shift towards higher precision image and video formats seems to be imminent. To overcome limitations of the common image and video formats, such as JPEG, PNG or MPEG, we propose two compression algorithms for High Dynamic Range (HDR) visual data: for static images and for video. The static image compression is intended for the OpenEXR format. They are both based on the well established standards, such as JPEG and MPEG, but use a novel color space, which can accommodate an extended dynamic range and guarantees the precision that is below the threshold of contrast detection. We show that only minor changes are required to the existing encoding algorithms to significantly enhance information content of the visual data. We argue that the HDR representation is a simple and universal way to encode visual data independently of the display or capture technology. We show usefulness of HDR data on the examples of post-processing effects. 6057-34, Poster Session Psychophysical measurement for perceptual image brightness enhancement based on image classification I. Kim, W. Choe, S. Lee, SAMSUNG Advanced Institute of Technology (South Korea) The purpose of this study is to examine the difference in perceptual brightness enhancement per image category through perceptual brightness measurement. Perceptual brightness is measured via psychophysical experiment and brightness enhancement is performed by TMF(Tone Mapping Function). The classification process is comprised of two steps. It is possible to classify histograms into six groups. The three different TMFs for each category selected using the criteria and TMF application strengths. A psychophysical experiment to measure perceptual brightness enhancement was carried out. The experiment was to determine the equal perceptual brightness point between an original image and the corresponding set of TMF images. The results showed that the mean luminance for each category is significantly different. The results from brightness matching indicate that perceptual brightness enhancement is dependent on image category. We can propose that image category should be considered for advanced brightness enhancement methods.
100
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6057: Human Vision and Electronic Imaging XI
6057-35, Poster Session Visual deficiency and image recognition: an image semantic cartography related to visual performance A. Scherlen, J. Da Rugna, Univ. Jean Monnet Saint-Etienne (France) In this study we test the benefit of an adaptive visual aid called VISAR (Visual Signal Adaptive Restitution) to help patients presenting visual impairment to recognize image. Actually the leading cause of vision disability is age-related macular degeneration (ARMD) affecting high-resolution vision. This pathology appears by the presence of a central scotoma masking the central part of the observed visual scene. Many aids are purposed but patients are still constrained to adapt to their scotomas by exploring and exploiting their remaining functional field of vision. This adaptation takes time, gives random results and requires great efforts on behalf of the patients. To limit this inappropriate behaviour our system will adapt in real time the image to the patient deficiency. By restoring pertinent semantic elements, hidden by the scotoma, to the patient visible zone we facilitate the image comprehension. At the opposite of actually aids, our concept takes into account patient cognitive behaviour during visual information integration. Finally, the results of some experiences show how this system may be useful for low vision rehabilitation. The deepening of the spatial reference integration in our system would improve patient performance during dynamics exploration. 6057-36, Poster Session Simple color conversion method to perceptible images for color vision deficiencies M. Meguro, C. Takahashi, T. Koga, Yamagata Univ. (Japan) In this paper, we propose a color conversion method for realizing barrier free systems for color-defective vision. Human beings perceive colors by a ratio of reaction values by three kinds of cones on the retina. The three cones have different sensitivity to a wavelength of light. Nevertheless, dichromats, who are lacking of one of the three cones, tends to be difficult for discriminating colors of a certain combination. The proposed techniques make new images by converting color for creating perceptible combination of color. The proposed method has three parts of processes. Firstly, we do image segmentation based on the color space L*a*b*. Secondly, we judge whether mean colors of divided regions of the segmented image tend to be confusion or not by using confusion color loci and color vision models of the persons with color-defective vision. Finally, the proposed technique realizes the perceptible images for dichromats by changing the confusion color in several regions of images. We show how effectiveness of the method by some application results. 6057-38, Poster Session Toward a taxonomy of textures for image retrieval J. S. Payne, Buckinghamshire Chilterns Univ. College (United Kingdom); T. J. Stonham, Brunel Univ. (United Kingdom) Image retrieval remains a difficult task, in spite of the many research efforts applied over the past decade or more, from IBM's QBIC onwards. Colour, texture and shape have all been used for content based image retrieval (CBIR); texture is particularly effective, alone or with colour. The Brodatz photographic album "Textures" is still widely used to provide a relatively small standard test set of images.
Many researchers have expressed the hope that textures can be organised and classified in the way that colour can; however, it seems likely that such an ambition is unrealisable. While the goal of content based retrieval is to retrieve "more images like this one", there is the difficulty of judging what is meant by similarity for images. It seems appropriate to search on what the images actually look like to potential users of such systems. No single computational method for textural classification matches human perceptual similarity judgements. However, since different methods are effective for different kinds of textures, a way of identifying or grouping such classes should lead to more effective retrievals. The perceptual similarity judgements from the experiments carried out in this research are applied to classify the Brodatz textures into eight perceptually matched groups. 6057-39, Poster Session Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity M. Phielipp, J. A. Black, Jr., S. Panchanathan, Arizona State Univ. To facilitate collaboration between computers and people, computers should be able to perceive the world in a manner that correlates well with human perception. A good example of this is face image retrieval. Mathematically-based face indexing methods that are not based primarily on how humans perceive faces can produce retrievals that are disappointing to human users. This raises the question "Can human faces be automatically indexed in a manner that correlates well with human perception of similarity?" Humans use words to describe faces - words such as bald, bearded, bespectacled, black-eyed, blond, blondish, blue-eyed, braided, brown-eyed, bucktoothed, dark-skinned, earringed, freckled, graybearded, or mustached. Such words represent dimensions that span a shared concept space for faces. Therefore they might provide a useful guide to indexing faces in an intuitive manner. This paper describes research that uses descriptive words such as these to index faces. Each word guides the design of one feature detector that produces a scalar coefficient, and those coefficients collectively define a feature vector for each face. Given these feature vectors, it is possible to compute a similarity measure between pairs of faces, and to compare that computed similarity to the similarity, as perceived by humans. 6057-40, Poster Session Subjective video quality evaluation for multimedia applications Q. Huynh-Thu, Psytechnics Ltd. (United Kingdom); M. Ghanbari, Univ. of Essex (United Kingdom); D. Hands, M. D. Brotherton, British Telecommunications plc (United Kingdom) Video quality can be measured using both subjective and objective assessment methods. Subjective experiments are important as they constitute a benchmark for evaluating the performance of objective quality metrics. Subjective quality assessment of television pictures has received extensive attention from experts over the past decades. On the other hand, emerging applications such as PCbased and mobile video streaming require new subjective test methodologies. Although some recent work has compared different methodologies and procedures, most concerned television pictures or were designed specifically for comparison of codec performance. No studies really assessed the repeatability and reliability of the experimental procedure. This paper outlines a methodology for conducting subjective evaluation of video quality for multimedia applications in a repeatable and reliable manner across different
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
101
Conf. 6057: Human Vision and Electronic Imaging XI
laboratories. Using video material at low-resolution, low-bit rate and low-frame rate, the same experiment was conducted by two different laboratories but set-up was slightly different, i.e. different computers and display panels were used, and viewing distance was not enforced to a fixed value. Results show that quality ratings obtained in both experiments are statistically identical. This is an important validation step for the Video Quality Experts Group, which will conduct an ambitious campaign of subjective experiments using many different test laboratories. 6057-41, Poster Session Texture segmentation using adaptive Gabor filters based on HVS S. Bi, D. Liang, Dalian Maritime Univ. (China) A texture segmentation algorithm based on HVS (Human Visual System) is proposed in this paper. Psychophysical and Neurophysiological conclusions have supported the hypothesis that the processing of afferent pictorial information in the HVS (the visual cortex in particular) involves two stages: the preattentive stage, and the focused attention stage. To simulate the preattentive stage of HVS, ring and wedge filtering methods are used to segment coarsely and the texture number in the input image is gotten. As texture is the repeating patterns of local variations in image intensity, we can use a part of the texture as the whole region representation. The inscribed squares in the coarse regions are transformed respectively to frequency domain and each spectrum is analyzed in detail. New texture measurements based on the Fourier spectrums are defined. Through analyzing the measurements of the texture, including repeatability directionality and regularity, we can extract the feature, and determine the parameters of the Gabor filter-bank. Then to simulate the focused attention stage of HVS, the determined Gabor filter-bank is used to filter the original input image to produce fine segmentation regions. This approach performs better in computational complexity and feature extraction than the fixed parameters and fixed stages Gabor filter-bank approaches. 6057-42, Poster Session Image stability analysis on the human retina M. I. Baritz, Univ. Transilvania din Brasov (Romania) In this paper it is presented some aspects about the studies on the human eye, the analyze and studies by simulation mechanism to establish the image stability on the human retina. In the first part we make an optical analyze on the transparent layers of the eye to obtain the image quality on the human retina. In these conditions, the most important part of the eyeball it's the lens, with its possibilities for accommodation during the visual process. In some cases a generalized behavior characterized by the optical transfer function (OTF) is sufficient, but more detailed model of the optical system (lens of the eye) is often needed, specially when we want to make a most effectual and better prosthesis for this part. The entire model of the eye considers an optical system formed by ten (10) centered quadratic refracting surfaces with rotational symmetry. The radius and asphericity define each surface. The lens of the eye is defined by radius of the first and final surfaces and its thickness on the axis, variable with accommodation. Because in the internal structure of the eye lens there are ten different surfaces, with different refractive index (different biomaterials) a model of this kind of lens have been computed by optical combination optimum method (OCOM). The relations between curvature radius, thickness of the eye lens and refractive index with value of accommodation are: where A = 0 - 10 m-1 To understand the exactly working mechanism of the entire eye in the accommodation process, the elastic behavior of the eye lens is
studied by Finite Element Method (FEM). The lens was considered as an axial-symmetrical structure, so, the problem becomes a 2D one. Five layers with different mechanical properties compose the lens and the external muscles were simulated by a force system having the resultant equal to zero. Initially the problem was considered geometrical non-linear one. And the dependence forcedisplacement resulted almost linear, therefore in the next runs a linear model was used. It was noticed that Poisson's ratio has an important influence on the elastic behavior of the eye lens; this aspect must be take into account for synthetic eye lens manufacturing. For understanding how we can obtain a quality image on the retina, it must analyze this complex optical system movement in the axis 3D system. From different medical observations on the visual stability, it can observe a very good concordance between the eye lens accommodation and the movements of the eyeball. Fig. 1. Fix system coordinates in the space of the surface (X,Y,Z) and the fix system on the eyeball (x,y,z) For that, each object point M(X,Y,Z) from fix system it is represented, in mobile system, by position vector r(t) like: (1.) where: r(t) = [x(t)/y(t)/z(t)]; and B(t) is orthogonal matrix 3x3 with its elements: functions a, b, g. (2.) The null value of the time variation of the vector r(t) will establish the stability conditions in the systems choosing. (3) (4.) For solve these equations we must choose three different points on the retina surface with the same illuminated function. We consider the illumination function E=E(X,Y,Z), invariable with the time, on observation surface and for the retinal surface another function: L=L(f, q, t). The receiver elements of the retina (cone and ......) transform this radiant energy into a potential U=U(q,f,t). The relation between L and U is a logarithmic one, but the variations are so little and we neglected the logarithmic variation and we were transform into linear relation U=kL(q,f,t). In this moment we can establish that the tension variation dU/dt on the two directions are the same, that mean the variations from object plan are formed on the surface of the retina in the same quantities. From this analyze and demonstration used for less three points in the object plan which are making the image on the retina surface, it can be obviously that there is a invariability of the making mode of the image on the human retina, even the eyeball is moving. Finally we make a simulation on the computer with these determinations of eyeball movement invariability. 6057-24, Session 7 A closer look at texture metrics H. H. Shenas, V. Interrante, Univ. of Minnesota This paper presents some insights into perceptual metrics for texture pattern categorization. An increasing number of researchers in the field of visualization are trying to exploit texture patterns to overcome the innate limitations of three dimensional color spaces. However, a comprehensive understanding of the most important features by which people group textures is essential for effective texture utilization in visualization. There have been a number of studies aiming at finding the perceptual dimensions of the texture. However, in order to use texture for multivariate visualization we need to first realize the circumstances under which each of these classification holds. In this paper we discuss the results of our three recent studies intended to gain greater insight into perceptual texture metrics. The first and second experiments investigate the role that orientation, scale and contrast play in characterizing a texture pattern. The third experiment is designed to understand the
102
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6057: Human Vision and Electronic Imaging XI
perceptual rules people utilize in arranging texture patterns based on the perceived directionality. Finally, in our last section we present our current effort in designing a computational method which orders the input textures based on directionality and explain its correlation with the human study. We will also present a discussion of human subjects' results. 6057-25, Session 7 M-HinTS: mimicking humans in texture sorting E. L. van den Broek, Vrije Univ. Amsterdam (Netherlands); T. Kok, T. E. Schouten, Radboud Univ. Nijmegen (Netherlands); E. M. van Rikxoort, Univ. Medisch Ctr. Utrecht (Netherlands) In human perception, computer vision, and image/video processing, texture analysis plays an important role. In this research, we analyzed human texture sorting and compared it with that of an artificial classifier (k-means clustering), using the VisTex and OuTex texture databases. Two experiments were done, in which 180 texture images were sorted. Both experiments were identical except for the fact that in one of them the color was removed from the texture images. For the experiments, an online experimental environment was developed: http://eidetic.ai.ru.nl/M-HinTS/ . The results of the experiments were compared with each other and with those of the artificial classifier that utilized the recently introduced scheme for parallel-sequential texture analysis. A range of analyzes were conducted, which revealed the interpersonal variability and identified the strategies groups of participants used. Moreover, differences in the sorting strategies between the color and gray scale texture images were found, which illustrates the importance of taking color into account when analyzing texture. No generic artificial texture sorting algorithm is introduced since humans themselves disagree; however, human texture sorting is mimicked for several groups of humans, each with their own strategies and characteristics. 6057-26, Session 7 Inference and segmentation in cortical processing Y. Liu, G. A. Cecchi, A. R. Rao, J. Kozloski, C. Peck, IBM Thomas J. Watson Research Ctr. We present in this paper a model of cortical architecture and processing aimed at accounting, in a biologically plausible way, for the inference and segmentation capabilities displayed by the visual system. The units of computation in the model represent small localized ensembles of interacting neurons, and their oscillatory dynamics is represented by amplitude and phase variables. The model's architecture is organized in a hierarchy, such that the receptive field size increases as the hierarchy goes up, but the amplitude and phase information are transmitted up and down the hierarchy. We show that our network, trained using a self-organized paradigm, can: (1) improve recognition robustness upon input degradation due to the effect of top-down information, and (2) solve the superposition problem; explicitly, this is done by clustering in different phases the components, throughout the network, of inputs that are presented simultaneously and recognized as distinct by the upper layers. We believe that this is a significant contribution towards understanding the problem of invariant recognition and segmentation problem of occluded objects by the visual cortex.
6057-28, Session 7 Perceptually based techniques for semantic image classification and retrieval D. Depalov, T. N. Pappas, Northwestern Univ. The accumulation of large collections of digital images has created the need for efficient and intelligent schemes for content-based image retrieval. Our goal is to organize the contents semantically, according to meaningful categories. We present a new approach for semantic classification that utilizes a recently proposed color-texture segmentation algorithm (by Chen et al.), which combines knowledge of human perception and signal characteristics to segment natural scenes into perceptually uniform regions. The color and texture features of these regions are used as medium level descriptors, based on which we extract semantic labels, first at the segment and then at the scene level. The segment features consist of spatial texture orientation information and color composition in terms of a limited number of locally adapted dominant colors. The focus of this paper is on region classification. We use a hierarchical vocabulary of segment labels that is consistent with those used in the NIST TRECVID 2003 development set. We test the approach on a database of 5000 segments obtained from 2000 photographs of natural scenes. For training and classification we use the Linear Discriminant Analysis (LDA) technique. We examine the performance of the algorithm (precision and recall rates) when different sets of features (e.g., one or two most dominant colors versus four quantized dominant colors) are used. 6057-30, Session 8 Symbol discriminability models for improved flight displays A. J. Ahumada, Jr., M. Trujillo San-Martin, J. Gille, NASA Ames Research Ctr. Computer models of human vision that predict human visual task performance can provide an objective means for improving flight display design by predicting pilot performance. An important step in the development of airborne traffic management displays is the evaluation of the discriminability of proposed symbol sets. Here we present an image discrimination model for symbol discriminability that includes both size and position compensation. The model takes as input the luminance values for the pixels of two symbol images, the effective viewing distance, and gives as output the discriminability in just-noticeable-differences (d'), the size reduction of the larger symbol, and the horizontal and vertical image offsets in pixels needed to minimize the discriminability. 6057-32, Session 8 Is haptic watermarking worth it? M. Barni, D. Prattichizzo, G. Menegaz, A. Formaglio, M. Franzini, Univ. degli Studi di Siena (Italy); H. Z. Tan, Purdue Univ. We propose a psychophysical experiment aiming at investigating the differential sensitivity of the visual and haptic channels in view of the design of a novel perception-driven watermarking technique. A very simple stimulus, composed by a flat surface with a watermark superimposed to it, is presented to the subjects that sense it either visually or haptically. A noise sensitivity paradigm is followed to estimate the detection threshold of the embedded watermark. The characterization of the differential sensitivity of the two perceptual channels in a typical visuo-haptic rendering set up is aimed at determining which of the two considered sensory channels (namely vision and touch) sets the boundary for the detectability of the watermark for given rendering conditions.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
103
Conf. 6057: Human Vision and Electronic Imaging XI 6057-33, Session 8 Display conditions that influence wayfinding in virtual environments R. A. Browse, D. W. S. Gray, Queen's Univ. (Canada) As virtual environments may be used in training and evaluation for critical real navigation tasks, it is important to investigate the factors influencing navigational performance in virtual environments. We have carried out controlled experiments involving two visual factors known to induce or sustain vection, the illusory perception of selfmotion. The first experiment had subjects navigate mazes with either a narrow or wide field of view. We measured the percentage of wrong turns, the total time taken for each attempt, and we examined subjects' drawings of the mazes. We found that a wide field of view can have a substantial effect on navigational abilities, even when the wide field of view does not offer any additional clues to the task, and really only provides a larger view of blank walls on the sides. The second experiment evaluated the effect of perspective accuracy in the scene by comparing the use of displays that were corrected for changing head position against those that were not corrected. The perspective corrections available through head-tracking did not appear have any influence on navigational abilities. Another component of our study suggests that during navigation in a virtual environment, memory for directions may not be as effective as it could be with supplemental symbolic representations.
104
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
Tuesday-Thursday 17-19 January 2006 Part of Proceedings of SPIE Vol. 6058 Color Imaging XI: Processing, Hardcopy, and Applications
6058-01, Session 1
Ideal illuminants for rod/L-cone color J. J. McCann, McCann Imaging In average moonlight natural objects appear colorless. In these cases there is only sufficient light to excite the rod-shaped receptors in human retina (scotopic vision). The rods are physiologically different from cone-shaped receptions as shown by many psychophysical experiments including: spectral sensitivity, the Stiles-Crawford effect, the flicker-fusion frequency, the shape of dark-adaptation curves, the apparent color and the apparent sharpness. All of these techniques have shown that complex color are seen by the interaction of rods with long-wave cones images at very low radiances, in illuminants rich in long-wave light. This paper measures the spectra of wood-fire and candle flames. Further, it shows that fire is the ideal illuminant for stimulating color with rods and L-cones at minimal radiances. In fact, complex color images are seen with radiances 100 to 1,000 times below M- and S-cone absolute thresholds. Spectral and DNA studies have shown that primates evolved Lcones in recent primate history. Anthropological studies have currently dated the controlled use of fire to 1.6 million years ago (Mya). The paper will discuss the evidence for dating the evolution of L-cones and the controlled use of fire. 6058-02, Session 1 Accuracy of color transforms P. J. Green, London College of Communication (United Kingdom) Colour management systems need to be accurate and repeatable, particularly in graphic arts applications such as converting image data to proofing systems and to final print processes. Modelling and computation errors were analyzed for a number of characterization data sets including single press sheet measurements and reference printing conditions based on SWOP and ISO 12647-3. Errors were calculated at double precision and compared with the AToB, BToA and round trip errors found in ICC profiles generated from the same data sets. Data sets which are averaged from measurements of multiple press sheets consistently performed better than transforms built from single-sheet measurements. A significant part of the transform error lies in the limited precision of computation. BToA transform errors averaging 1 + E*ab appear to be a reasonable expectation for well-formed data sets. This needs to be placed in the context of errors arising from measurement uncertainty and the variation in colour reproduction systems, which are considerably larger. 6058-03, Session 1 Color image dequantization by constrained diffusion D. Keysers, T. M. Breuel, DFKI GmbH (Germany) and Univ. of Kaiserslautern (Germany) We propose a simple and effective method for the dequantization of color images, effectively interpolating the colors from quantized levels to a continuous range of brightness values. The method is designed to be applied to images that either have undergone a manipulation like image brightness adjustment, or are going to be processed in such a way. Such operations often cause noticeable color bands in the images. These artifacts can be reduced using the proposed Constrained Diffusion technique. We use smoothing by isotropic diffusion and constrain the maximum difference to the original image in each step of the diffusion process.
If used before the manipulation, we change the image representation from 8 to 16 bits per channel and apply the diffusion while restricting the maximum change to the width of the quantization interval of the 8 bit values in 16 bit. Thus, the algorithm respects the constraints implied by the original 8 bit data. We demonstrate the advantages of our method using synthetic and real life images as examples. We also present quantitative results using 8 bit data that has been obtained from original 12 bit sensor data and obtain substantial gains in PSNR using the proposed method. 6058-04, Session 1 Spring-primary mapping: a fast color mapping method for primary adjustment and gamut mapping H. Zeng, Hewlett-Packard Co. A gamut mapping method, spring-primary gamut mapping, was developed for device to device color mapping. Instead of performing gamut mapping point-wisely, it selects a small number of the nodes of a lookup table for gamut mapping and determines the color mapping of other nodes by interpolation using color similarity information. Primary adjustment is seamlessly incorporated into the gamut mapping step, therefore it further improves the performance. Because the color similarity information of neighbor colors are used for the primary adjustment and gamut mapping, the color to color relative relationship is well retained. The color mapping for gamut surface colors can be configured for the preference of business graphics or for the tradeoff of business graphics and pictorials. As the processing goes from gamut surface colors to interior colors, the color mapping is gradually adapted to the preference of pictorials. With this unique property, a color map produced by this method can be used for the color transformation of both photographic images and business graphics. Because gamut mapping is performed only on a small percentage of nodes, the color mapping is highly efficient. 6058-05, Session 1 A framework for image-dependent gamut mapping J. Giesen, E. Schuberth, ETH Zьrich (Switzerland); K. Simon, EMPA (Switzerland); D. Zeiter, ETH Zьrich (Switzerland); P. Zolliker, EMPA (Switzerland) Gamut mapping is a multi-criteria optimization problem with several competing objectives like target gamut exploitation, continuity, contrast-, saturation-, gray axis- and hue preservation. Depending on the application any of these objectives is more or less important. The scenario that we want to address here is image dependent gamut mapping. Since image gamuts can be small and odd shaped of-the-shelf gamut mapping algorithms either cannot be used in this scenario or often result in far from optimal target gamut exploitation. In our approach we formulate image dependent gamut mapping as a parameterized mathematical optimization problem that allows constraining the degree to which objectives like contrast preservation, hue preservation, saturation preservation and the continuity of the mapping can be violated while maximizing the target gamut exploitation. The resulting optimization problem is in a general non convex form and thus not efficiently solvable in practice. But slight modifications of the optimization problem turn it into a convex quadratic program, which can be solved efficiently in
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
105
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
practice. We demonstrate the feasibility of our approach on several benchmark image- and device gamuts. 6058-06, Session 1 Perceptual gamut mapping algorithm development based upon image quality and preference factors B. Kang, M. Cho, H. Choh, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) This study has three primary aims circumventing current limitations of color reproduction technologies: firstly, to derive base-line image quality factors from both color printer experts and academic research works. Individual factors were verified by systematic experiments, secondly, to develop a perceptual gamut mapping algorithm covering the image quality and preference factors derived, thirdly, to apply the algorithm to printer driver as acting for a vendor specific perceptual intent. Algorithm of this study tried to optimization between control parameters of gamut mapping and color shifting factors of preference, e.g. skin, sky and green grass. Profile builder using this algorithm outperforms, in industrial and academic aspects, existing commercial tool and CIE recommended algorithms. 6058-07, Session 1 Gamut estimation using 2D surface splines M. Q. Shaw, Hewlett-Packard Co. This paper proposes a method of gamut estimation using data segmentation and 2D surface splines. The device data is first segmented into hue intervals, and then each hue interval is analyzed iteratively to construct a 2D gamut boundary descriptor for that hue interval. The accuracy and smoothness of the gamut boundary estimate can be controlled by the number of hue intervals selected and the sampling within each hue interval. 6058-40, Poster Session High-resolution image viewing on multiprojector distributed parallel rendering display wall J. Meng, H. Lin, J. Shi, Zhejiang Univ. (China) We illustrate an application that displays high-resolution images and play panoramas on Multi-Projector Tiled Display Wall, such as photos from satellites and microscopes, or landscape sceneries. Our prototype is implemented over a local area network with a PC cluster and a monitor. Image data are initially stored in the monitor PC, and then dispatched to the displayer PCs and is projected on to the display wall. In this paper we focus on three main topics: 1) We use a single texture mapped on to each displayer's screen. The texture is from a high resolution image which exceeds the capacity of GPU. To overcome this problem without reducing the resolution, we use a multi-resolution texture hierarchy, and select one of them dynamically depending on which part of the image is requested. 2) We use binary-tree transferring structure to reduce the latency from O(N) to O(log2 N)(N is the number of Displayer PCs). 3) To playing the panorama, we use parallel Macro-Micro clocks for system synchronization in order to eliminate accumulated displacement and in the mean time to achieve a high frame-rate. The result with a 10548*15474 image is shown in the paper.
6058-41, Poster Session The application of wavelet transforms and mathematics morphology on the processing of infrared satellite cloud image J. Xue, Z. Liu, P. Wang, Tianjin Univ. (China) Image segmentation is an important and fundamental task in many digital image processing systems and computer vision. The aim of the intelligent weather forecast system is to forecast disastrous weather such as typhoon and storm exactly. Infrared satellite cloud images (ISCI) are important resources of the system. It is difficult to segment and detect the boundary of cloud images because of the complicated and various shapes and blurry edges of cloud. In this paper, we present an idea and a set of realizable design about selfadaptive segmentation by means of mathematical morphology firstly, which can segment the useful clouds that are larger and have lower temperature from the cloud image. Then, we proposed a boundary detection method base on wavelet transform. This boundary detecting method of the infrared satellite cloud image, which is based on Bubble wavelet function and Gaussian function, is suitable for dealing with the neighborhood processing. Some practices show that the segmentation models are self-adaptive, the whole processing system is general and the algorithm is high efficient in the processing of infrared satellite cloud image. 6058-42, Poster Session Simulation and parameter optimizing of multielectrode capacitive transducers based on finite element method D. Chen, Harbin Univ. of Science and Technology (China) The principle of electrical capacitance tomography (ECT) is introduced and the simulation based on the finite element method (FEM) about the sensitivity distribution of 12-electrode capacitance transducers for oil and water two-phase flow system is discussed. Unequal mesh spacings in cross section of Transducers are adopted in automatic plotting to improve the precision in calculation. The optimal function is presented based on sensitivity distribution which plays an important role in image reconstruction. The influence which the structural parameters of transducers impose on the performance of ECT is studied. EC changes of transducers between empty pipeline and full pipeline can express the global sensitivity of all elements for transducers in the pipeline, therefore EC changing quantity between empty pipeline and full pipeline can be considered as an important index of performance for transducers. A FE simulation analysis is done on performance of transducers which have different structural parameters, and the main analysis is focused on influences that changes of various structural parameters exert on EC and its changing quantity between empty pipeline and full pipeline. The parameter optimizing is achieved by employment orthogonal design method combined with FEM. 6058-43, Poster Session Digital watermarking of color image S. C. Chao, Ta Hwa Institute of Technology (Taiwan); H. M. Huang, Tung Nan Institute of Technology (Taiwan); C. Y. Chen, Institute of Nuclear Energy Research (Taiwan) An analytic method of embedding color watermarks into the color images is proposed. The original image is color image of size 480X640, and the watermark image is 24-bit-gray scale image of size 120X160. The quality variations of the embedded image and extracted watermark are analyzed. From the experimental results they appear that the quality of embedded image varies with the quantization index. They also show that the quality of extracted watermark is dependent on the gray-level of the original image. The relationship between the gray-level and the noise of embedded
106
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
image has been quantitatively analyzed. The digital quantization method is also important to decrease the noise of embedding image and the error of extracting watermark. 6058-45, Poster Session Subjective assessment of printed color image quality as saturation of the primary colors RGB decreases W. Song, H. Seki, Naltec Inc. (Japan); G. Ohashi, Y. Shimodaira, Shizuoka Univ. (Japan) This study is to obtain the quantitative data of the quality of printed images as a function of the saturation decrease of three primary colors. As the results, decrease of color saturation of three primaries significantly affects the quality of printed images. The saturation decrease of primary B has the most severe influence on the quality impairment of the printed images. According to analysis the results, it is found that the values of the saturation decrease ratio, corresponding to the acceptable limits of image quality, are 13% for primary R, 11% for primary G, and 1% for primary B by proceeding the assessment data which is obtained from the single-stimulus method. Applying these saturation decrease values in CIELAB a*b* plane, the value of gamut area ratio of primary R, G, and B, corresponding to the acceptable limits of image quality, are 70%, 80%, and 96% of that of original image gamut. Furthermore, the reliability of the saturation decrease ratios from single-stimulus method is verified by using the paired-comparison method. The results show a way for estimating the amount of saturation decrease according to the quantitative data of quality of a printed image and indicate a capability to control the saturation decrease of three primary colors to keep the printed images quality constantly. 6058-08, Session 2 Uncalibrated color N. Moroney, Hewlett-Packard Co. No abstract available 6058-09, Session 2 Geometrical methods for lightness adjustment in YCC color spaces R. Samadani, G. Li, Hewlett-Packard Labs. Lightening or darkening an image is a fundamental adjustment used to improve aesthetics or correct exposure. This paper describes new geometrical algorithms for lightness adjustment, implementing fast traversal of colors along lightness-saturation curves, applicable when the data starts naturally in YCC space (JPEG images or MPEG videos). Here, YCC refers generically to color spaces with one luminance and two color difference channels, including linear YCC spaces and CIELAB. Our first solution uses a class of curves that allows closed-form computation. Further assuming that saturation is a separable function of luminance and curve parameter simplifies the computations. This approach reduces clipping and better adjusts lightness together with saturation. Preliminary evaluation with 96 images finds good subjective results, and clipping is reduced to about 5% of a prior approach. 6058-10, Session 3 Measuring gloss by digital photography P. Kumar, L. W. MacDonald, London College of Communication (United Kingdom) Gloss is an important attribute of appearance. The most common visual characteristic of a glossy object is that it shows bright highlight details where the incident illumination is reflected in the
specular direction. The measurement of gloss is conventionally made by specialised instruments that determine the ratio of reflected to incident illumination at a single fixed angle. This presentation will describe a study that investigated whether digital photography with flash illumination could be used as an alternative. Multiple exposures were combined by a high dynamic range (HDR) imaging technique to produce a two-dimensional intensity profile of the reflectance around the specular point. The method was tested for six paper samples of varying gloss, and the results were found to correlate well with instrumental measurements. The image gloss profiles, however, provided more information about the distribution of the reflection over a range of angles and also gave an indication of the surface texture. 6058-11, Session 3 Ubiquitous image processing: a novel image-enhancement facility for consumers R. Shaw, P. Johnson, White Rose Digital While the number of digital photographs is still growing exponentially, experience shows that more than fifty-per-cent of all acquired pictures could be significantly enhanced for increased consumer satisfaction. Thus, while the market potential for userfriendly image-enhancement software is huge, most existing imageenhancement software is either over-simplistic, user-hostile, or both. The authors will describe a novel approach to this practical problem, based on the definition of a fundamental set of physical vectors representing any digital image. These independent vectors are then placed within an overall logical hierarchy, and practical linear visual ranges are established for a comprehensive sample of digital color images. Finally increments are set within these ranges in terms of just-appreciable visual differences. This methodology typically produces a hierarchy of separate image-states of the order of a hundred thousand or more for a typical digital photograph, and the problem remains of presenting these to the user for personal choice of optimum enhanced image-quality. By implementation of a critical comparative hierarchy of image choices, itself based on an information-theoretic sorting-matrix, the user is presented with a relatively small number of images in navigating through very this large number of independent image states. Software has been developed whereby an unskilled user may enhance and optimize the image quality of any digital photograph to personal choice in a matter of several seconds, and, based on the manner in which the independent vectors are defined, without the introduction of any unwanted image artifacts that are typical of many software packages. A practical demonstration will be given of this software, which by its simple user-interface, real-time computation, and lack of any user learning-curve, naturally lends itself to many practical imaging applications in addition to a stand-alone software package, including digital cameras, printers, scanners and photo-kiosks, or provision as a web-service. 6058-12, Session 3 Color constancy on Japanese animation Y. G. Ichihara, Hosen-Gakuen College (Japan) This paper focuses on Imagined Color System in Japanese animation. We measured the color constancy and luminance of two scenes from a Japanese animation fil m. The data show that the color system used in the film is not natural but an imagine and artistic appearance system. Color constancy of human vision can te ll the difference in skin and hair colors between under moonlight and day ligh t. Human brain generates a match to the memorized color of an object from dayl ight viewing conditions to the color of the object in different viewing condit ions. For example, Japanese people always perceive the color of the Rising Sun in the Japanese flag as red even in a different viewing condition such as und er moonlight. Color images captured by a camera cannot present those human per ceptions. However, Japanese colorists in Animation
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
107
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
succeeded in painting the e ffects of color constancy not only under moonlight but also added the memory m atching colors to increase the effect of the scenes on the viewers` perception s. Its importance is that it could also provide an explanation on how human br ain perceives the same color under different viewing conditions. 6058-13, Session 3 Convincing non-printers to become future customers R. Fageth, W. Schmidt-Sacht, CeWe Color AG & Co. OHG (Germany) The number of images taken by digital still cameras (DSC) and camera phones is rising dramatically but the number of digital images being transferred onto paper is not keeping up with that rate of increase. This paper tries to evaluate why this is the case and offers suggestions on how to encourage consumers to print their most valuable memories while using convenient solutions both at home and professionally. Paths from a digital file to a print are evaluated and the diversity of possibilities for the consumer are described. An approach to offer the consumers all possibilities of products from classical silver halide prints up to semi-professional photo books printed on electro photography digital presses is presented. This approach offers also the possibility - compared to classical approaches of transferring PDFs from the client to the digital press to optimize every image separately before printing. 6058-14, Session 4 DTV color and image processing: past, present, and future C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) No abstract available 6058-15, Session 4 Subpixel rendering method for color error minimization on subpixel structured display W. Choe, S. Lee, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) This study investigates the color error problem posed by large flat panel displays and proposes a subpixel-rendering algorithm to mitigate the problem. On large subpixel structured displays, the Mach band effect and the convergence error of pixels make the color error, the color band error. The proposed method to reduce the color band error includes three processes: a process of finding the areas or pixels generating the error, an error-estimation process, and an error-correction process. To correct the color band error, we take an error erosion approach, an error concealment approach, and a hybrid of the erosion and concealment approaches. In this paper, we devised a psychophysical method to determine the detection threshold of color band error by the human vision. In addition, we applied our proposed method to a commercial 42" plasma display to confirm the effects. The results show that all observers see the color band error at sharp edges having above 64gray difference and the corrected test images by our algorithm are preferred to the original test images. Finally, this paper reports that the Mach band effect and the convergence error on large subpixel structured displays produce color band errors on images having sharp edges and the proposed method effectively corrects the color band errors.
6058-16, Session 4 Compensation method for color defects in PDP due to different time responses of phosphors H. Oh, H. Lee, D. Park, S. Kim, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) On a plasma display panel (PDP), luminous elements of red, green, and blue have different time responses. This difference generates a greenish colored trail and bluish edge behind and in front of a bright object moving on a dark background, respectively, or the opposite. Due to such phosphor lag effect, the quality of an image sequence deteriorates inevitably. The human eyes are usually less sensitive to gray than color variation on moving images. Therefore, in order to reduce the color artifacts, this paper proposes a motion-based discoloring method. To compensate for greenish trails, the video values of the blue and red sub-pixels located behind the object are increased artificially, based on the motion magnitude and direction, to match the afterimage of the green sub-pixel. For bluish edges, the values of the blue and red sub-pixels in front of the object are decreased artificially. Discoloring values are modeled as linear functions of a motion vector to reduce hardware complexity. Experimental results show that the proposed method has effectively removed the colored trails and edges of moving objects. Moreover, the clear image sequences have been observed compared to the conventional ones. 6058-17, Session 4 Six-primary-color LCD monitor using sixcolor LEDs with an accurate calibration system H. Sugiura, H. Kaneko, S. Kagawa, J. Someya, H. Tanizoe, Mitsubishi Electric Corp. (Japan) The authors have successfully developed a six-primary-color liquid crystal display using six-color LEDs. More specifically, a prototype of a six-primary-color liquid crystal display having a color gamut of 170% or wider then that of conventional techniques has been constructed. In order to obtain highly stable image quality for these high performance monitors with a wide color gamut, we considered the fluctuation of white balance. Some fluctuation factors are listed below, (1) Temperature dependency of luminance intensity of LEDs, (2) Temperature dependency of emission wavelength of LEDs, (3) Degradation of luminance intensity resulting from variations per hour The key factor in temperature change is an LED's own generation of heat. When the LEDs temperature changes, the luminescence intensity and the dominant wavelength of the LEDs are also changed accordingly, which have a severe effect on the color reproduction characteristics of the monitor. These LEDs characteristics lead to the fluctuation of the white balance of the backlight. This monitor has a newly developed calibration system using an integrated color sensor, so that it can keep its white point chromaticity stable. This research was organized in part by the New Energy and Industrial Technology Development Organization, Japan (NEDO). 6058-18, Session 5 A color control method for image output with projection displays S. Tominaga, K. Kumamoto, Osaka Electro-Communication Univ. (Japan) The present paper proposes a nonlinear approach using a neural network for color control of projection displays, including the LCD and DLP types. This approach accepts variations in primary color coordinates and coupling among RGBW channels. We regard a
108
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
display system as an unknown nonlinear system with RGB signal inputs and XYZ tristimulus outputs. We determine the RGB values so that the projector outputs the desired XYZ values. The neural network is used for estimating adaptively an inverse mapping from the XYZ space to the RGB space. Because of a direct mapping, we can eliminate the need to predict white channel separation. Moreover, we present a method for correcting the emitted luminance, according to the spatial location and the surface color of a screen. The spatial correction is needed because a color image from a computer is not displayed uniformly on the screen. The screen color correction makes it possible to reproduce accurate color images on a colored wall. 6058-19, Session 5 Illuminant-adaptive color reproduction for a mobile display J. M. Kim, K. Park, M. Lee, Y. Cho, Y. Ha, Kyungpook National Univ. (South Korea) Mobile displays, such as PDAs and cellular phones, are viewed under various lighting conditions. In particular, images displayed in mobile under outdoor environment are perceived as quite a bit dark due to the light adaptation of the human visual system. In addition, flare phenomena decrease the color gamut of a mobile display by increasing the luminance of black level and de-saturating. Therefore, this paper presents an illuminant adaptive reproduction method composed of lightness enhancement and chroma compensation. First, the ambient light intensity is measured using a lux-sensor, then the flare is calculated based on the reflection ratio of the display device and the ambient light intensity. Second, lightness is enhanced by linearization of the human's perceived lightness for input digital values, since the relative cone response in human eye is nonlinear to luminance of the scene. Also, lightness enhancement is considered that cone response vary according to ambient light intensity. Next, chroma of the displayed image is reduced by ambient light. Then physically reduced chroma between original image's chroma and added flare image's chroma is compensated for adding chroma difference depending on gamut boundary. Consequently, the proposed algorithm improves the quality of the perceived image adaptive to an outdoor environment. 6058-20, Session 5 Skin color reproduction algorithm for portrait images shown on the mobile display Y. Kwak, S. Lee, D. Park, C. Kim, SAMSUNG Advanced Institute of Technology (South Korea) The preferred skin color reproduction algorithm is developed for the mobile display especially for a portrait image with one person as a main object occupying most of the screen. According to the developed technique, the skin area in an image is detected using color value of each pixel in YCbCr color space. The skin color boundary is defined as a quadrangle in Cb-Cr plane. The colors of pixels belonging to skin area are shifted toward the preferred colors while there is no color change for the other pixels. The psychophysical experiments are conducted to investigate the optimal model parameters providing the most pleasant image to the users. Then, the performance of developed algorithm is tested using the optimal parameters. The result shows that for more than 95% cases, the observers prefer the images treated with the developed algorithm compared to the original image. It is believed that the developed algorithm can be applied to the mobile application to improve the image quality regardless the input sources.
6058-21, Session 5 Estimating displays' color fidelity based on classified image statistics P. Sun, C. Lee, Shih Hsin Univ. (Taiwan) The color fidelity of displays such as LCD and PDP was commonly estimated by a series of color patches. However, no strong evidence showed that these color patches are well correlated to the real-world image characteristics. To minimize the gap between the color patches and the image characteristics, the present study first accumulated the statistics of different classes of image. Then, we chose several of them as reference images, manipulated their colors intentionally to simulate uncalibrated displays, and finally asked observers to compare the reference images with their colorperturbed counterparts. In the end, based on the classified image statistics, an empirical model was derived to predict the overall results of the visual image differences. In the model, the pixel frequencies (probability) of 125 color clusters summarized by a certain type of images were taken into account for predicting the displays' color fidelity on the type of images. The 125 color clusters were linked in a 5x5x5 cellular structure. If the manipulation shifts the colors without disturbing the links, the reproduction still looks fine. Our results showed the variations of the links can be used as weights to enhance the performance of our model. 6058-22, Session 5 End-user display calibration via support vector regression B. Bastani, Hewlett-Packard Co.; B. V. Funt, W. Xiong, Simon Fraser Univ. (Canada) The technique of support vector regression (SVR) is applied to the color display calibration problem. Given a set of training data, SVR estimates a continuous-valued function encoding the fundamental interrelation between a given input and its corresponding output. This mapping can then be used to find an output value for a given input value not in the training data set. Here, SVR is applied directly to the display's non-linearized RGB digital input values to predict output CIELAB values. In comparison to some of the existing linear methods for calibrating different display technologies, including the gain-offset-gamma (GOG) and Masking Models, an advantage of using SVR for color calibration is that the end-user does not need to apply a different calibration model for each different display technology. We show that the same model can be used to calibrate CRT, LCD and DLP displays accurately. 6058-23, Session 5 The calibration accuracy of display white point by visual calibrator under various illuminations T. Sugiyama, Y. Kudo, Dai Nippon Printing Co., Ltd. (Japan) We have developed a method to calibrate a display to a predetermined state and to make an ICC display profile by visual calibration. Our method adjusts a color of display white point to that of a white object under a viewing illuminant, so our method can calibrate the display white point for any viewing illuminant. In this paper, we evaluated the matching accuracy between display white point and paper white under various illuminances and color temperatures. We found that the matching accuracy was almost the same in all illumination conditions. Furthermore, all subjects calibrated the chromaticity of display white point more bluish than that of paper white in all illumination conditions, and the dispersion of yellow-blue direction was larger than that of red-green direction. This yellow-blue direction was almost the same as the long axis of MacAdam ellipse. We also evaluated the capability of color
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
109
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
discrimination between display and paper. At 500lx or less, many subjects judged color difference correctly, though it was difficult to judge color difference correctly at 1000lx or more. These results suggest that ISO 3664:2000 P2 condition is appropriate to compare the image on computer display with that on paper at the same time. 6058-24, Session 6 Black extraction method using gamut boundary descriptors M. Cho, B. Kang, H. Choh, SAMSUNG Advanced Institute of Technology (South Korea) Color data conversion between CMYK and CIEL*a*b* color space is not directly corresponded, that is many CMYK combinations could reproduce the same CIEL*a*b* value. When building a LUT converting from CIEL*a*b* to CMYK for a CMYK color printer, one to one correspondence between CMYK and CIEL*a*b* must be aimed. The proposed method in this paper follows steps: (1) print and measure CIEL*a*b* values of CMYK reference chart, (2) set-up parameters to assign the amount of black extraction, (3) generate gamut boundary descriptors for gamut mapping and for black extraction using CMYK-CIEL*a*b* data under predetermined black extraction parameters, (4) perform gamut mapping for given CIEL*a*b* using the gamut boundary descriptor for gamut mapping, (5) determine K value of the gamut-mapped CIEL*a*b* using the gamut boundary descriptors for black extraction. The suggested method determines K value for given CIEL*a*b* using gamut boundary descriptors in CIEL*a*b color space. As a result, a color printer using this method can make out accurate black amount and reproduces more consistent CMYK images under different black extraction options. 6058-25, Session 6 Colorimetric characterization based on color correlation in CMYKGO printer I. Jang, C. Son, T. Park, K. Ko, Y. Ha, Kyungpook National Univ. (South Korea) This paper proposes a method of colorimetric characterization based on the color correlation between color patches in a CMYKGO printer. Many color patches represents the same tri-stimulus value, in colorimetric characterization beyond three colorants. Therefore, choosing the proper color patches corresponding to each tristimulus value is important for a CMYKGO printer characterization process. As such, the proposed method estimates the CIELAB value for many color patches, then selects certain color patches while considering high fidelity and the extension of the gamut. The selection method is divided into two steps. First, color patches are selected based on their global correlation, i.e. their relation to seed patches on the gray axis. Second, if the correlation factor is smaller than the correlation factors for neighboring patches, the color patch is reselected by new seed patch which is the average distribution of eight neighboring selected color patches. The selected color patches are then measured for accuracy, and the relation between the digital value and the tristimulus value for the color patches stored in a lookup table. As a result of this characterization, the gamut is extended in the dark regions and the color difference reduced compared to conventional characterization methods. 6058-26, Session 6 Hardcopy global color correction Y. Bang, Y. Kim, H. Choh, Samsung Advanced Institute of Technology (South Korea) As time, temperature or an external environment changes, a laser electrophotographic printer produces quite different color tones from
original ones. To achieve consistent color reproduction, many researchers have tried to characterize printer tone curves and developed methods to correct color tones. Color channel independent methods are most widely used, and there are two approaches in color channel independent method: (1) Instrumentbased correction and (2) visual correction. Two approaches provide some trade-offs between cost and accuracy. In this paper we propose a methodology which combines the strengths of these two approaches. We describe how we design a calibration page and how we characterize lightness variation of a reference patch. We then present the procedure of our global tone correction method based on visual appearance match of end-users as well as the predetermined reference lightness model. We simulate tone distortion state by varying hardware parameters, and perform visual appearance match experiments to subjects. Our experimental results show that our method can significantly reduce color difference between the original print and the print at the distortion state. This suggests that we can reliably estimate the distortion parameter, and correct tones close to an original state. 6058-28, Session 6 Efficient document rendering with enhanced run length encoding G. Feng, Ricoh Innovations, Inc.; C. A. Bouman, Purdue Univ. Document imaging and transmission systems (typically MFPs) require both effective and efficient image rendering methods that support standard data formats for a variety of document types, and allow for real time implementation. Since most conventional raster formats (e. g. TIFF, PDF and JPEG) are designed for use with either black and white text, or continuous-tone images, more specialized rendering methods are often required for representing mixed content documents. The baseline TIFF format supports a few binary compression options: PackBits, CCITT G3 and G4. Conventionally, halftoning algorithms, such as error diffusion, can be used to create a binary representation of a document image in the TIFF format. However, PackBits, CCITT G3/G4 compression generally do not produce desired compression on halftone images. In this paper, we propose an efficient error diffusion algorithm optimized for PackBits compression. This method, called POED (PackBits optimized error diffusion), is a form of threshold modulation error diffusion which takes advantage of the byte-oriented run length structure of PackBits compression by encouraging repetition of bytes in the binary output. To maintain the sharpness of text, a binary segmentation algorithm is used to switch on Floyd Steinberg error diffusion in text regions. The POED method yields higher PackBits compression ratios than conventional Floyd Steinberg error diffusion, while maintaining desirable visual quality with low computational and memory requirements. 6058-29, Session 7 Model-based clustered-dot screening S. H. Kim, SAMSUNG Electronics Co., Ltd. (South Korea) I propose a halftone screen design method based on a human visual system model and the characteristics of the electro-photographic (EP) printer engine. Generally, screen design methods based on human visual models produce dispersed-dot type screens while design methods considering EP printer characteristics generate clustered-dot type screens. In this paper, I propose a cost function balancing the conflicting characteristics of the human visual system and the printer. By minimizing the obtained cost function, I design a model-based clustered-dot screen using a modified direct binary search algorithm. Experimental results demonstrate the superior quality of the model-based clustered-dot screen compared to a conventional clustered-dot screen.
110
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
6058-30, Session 7 AM-FM hybrid color screen design to reduce brightness variation B. M. Kang, B. T. Ryu, C. Kim, Inha Univ. (South Korea); S. H. Kim, SAMSUNG Electronics Co., Ltd. (South Korea) AM-FM hybrid screen represents a clustered halftone screen whose centers of dot clusters are aperiodic and dot growth pattern is irregular. Compared to the AM ordered screen, AM-FM hybrid screen is free of inter-screen and subjective moirйs. However, it results in brightness variation often called as stochastic noise. In this paper, a new screen design technique is presented for the AM-FM hybrid dithering. Centers of dot clusters for each of C, M, Y, K screens are selected by a new method to achieve homogeneous distributions, channel by channel and as combination of color channels. An optimum dot growth filter that is different from donut filter is defined in this paper. Also, a new dot growth algorithm is developed to reduce the brightness variation. 6058-31, Session 7 Frequency domain design of cluster dot screens M. Fischer, D. Shaked, Hewlett-Packard Labs. (Israel) Clustered dots screens are widely used in digital printing. Our research focuses on irregular cluster-dot screens. Irregular screens are appealing since there are many more irregular screens than regular ones. As a result, they provide a larger set of multiseparation screen combinations for avoiding interaction between screening frequencies of concurrent screens. Their down side is that they often have harmonic artifacts. Our design challenge is to reduce these artifacts' visual impact. We present frequency-domain based methods to reduce these artifacts' visual impact. State-ofthe-art screens incorporate many, predominantly spatial domain, design considerations which we cannot ignore. Accordingly, the proposed methods are designed to introduce minimal modifications to given screens. The proposed methods eliminate, or reduce the few artifacts visible in a set of irregular screens. This work can be generalized to other printing technologies, and to screen design. 6058-32, Session 7 A spatial domain optimization method to generate plan-dependent masks Y. Wu, Hewlett-Packard Co. Stochastic screening technique uses a fixed threshold array to generate halftoned images. When this technique is applied to color images, an important problem is how to generate the masks for different color planes. Ideally, a set of plane dependent color masks should have the following characteristics: a) when total ink coverage is less than 100%, no dots in different colors should overlap from each other. b) for each individual mask, dot distribution should be uniform, c) no visual artifact should be visible due to the low frequency patterns. In this paper, we propose a novel color mask generation method in which the optimal dot placement is searched directly in spatial domain. The advantage of using the spatial domain approach is that we can control directly the dot uniformity during the optimization, and we can also cope with the color plane-dependency by introducing some inter-plane constraints. We will show that using this method, we can generate plane dependent color masks with the characteristics mentioned above.
6058-33, Session 8 Using errors in halftoning to increase reproduction accuracy S. Herron, Global Graphics Software Inc. Halftones are point-to-point conversion processes. This solution achieves accurate halftone reproduction by using the errors during the scan conversion process without comparison to the source pixels maintaining the point-to-point scan conversion paradigm. Two types of density errors encounter in the halftoning process are considered. An error describing the difference between the input pixel value and the halftone threshold value and an error describing the difference between the density level of the output and the reflection density of the tone. Halftone errors are metadata derived from a contone characteristic and communicated to the halftone conversion process. Each error is associated with a device dot maintaining the point-to-point relationship. An approach to achieving accurate halftone reproduction by using the errors is described. The algorithm is a two-pass process. During the first pass through the image, each pixel value in the image is converted to the threshold value according to the standard scan-conversion method. The difference between the threshold value and the vector or pixel value accumulates. That value is added to the average density increase. The second pass compares the value in the alpha plane with the threshold value. An adjustment is made to the halftone threshold value. 6058-35, Session 8 Analysis of misregistration-induced color shifts in the superposition of periodic screens B. Oztan, G. Sharma, Univ. of Rochester; R. P. Loce, Xerox Corp. We present an analysis and model for evaluation of color shifts in halftone printing caused by inter-separation misregistration for periodic clustered dot halftones. Using a lattice framework, we present intuitive analysis that demonstrates conditions under which the average color is asymptotically invariant under inter-separation misregistration. Combining the framework with an analytic representation for the halftone dots, we develop a hybrid analytical-numerical model for quantitatively estimating color shifts as a function of inter-separation misregistration. The model is compared against experimental data for a xerographic printer. 6058-36, Session 8 Analysis of color error diffusion with vector error filters Z. Z. Fan, Xerox Corp. As vector error filters are capable of passing errors generated in one color component to other color components, it provides more flexibility in shaping the halftone texture. As a result, it may potentially produce halftones with better image quality. In this paper, we analyze color error diffusion with vector error filters. In particular, we will discuss its halftone spectrum features and its stability conditions with respect to the filter coefficients. For spectrum analysis, we will derive the high-pass and "zero-gain" conditions, which ensure decent image quality. Since error diffusion is a feedback system, the vector error filters may cause instability, if it is not properly designed. This may potentially generate ever-increasing quantization error that masks the input and produces unacceptable output images. The stability conditions we will discuss prevent any instability.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
111
Conf. 6058: Color Imaging XI: Processing, Hardcopy, and Applications
6058-37, Session 8 New halftoning method combining the best of masking and error diffusion algorithms F. Cittadini, Ocй Print Logic Technologies (France) and Univ. Pierre et Marie Curie (France); J. Pervillй, S. Berche, Ocй Print Logic Technologies (France); M. Ben Chouikha, G. Alquiй, Univ. Pierre et Marie Curie (France) There are two main families among the halftoning methods: halftoning by masking (i.e. blue noise masking) and error diffusion halftoning. The first family produces neither "worms" nor defects related to stationary regimes but has a limited spatial bandwidth. The error diffusion halftoning method is characterized by a very broad spatial bandwidth allowing good rendition of very thin lines and patterns but this method presents sometimes unpleasant worms or stationary regimes. These methods are complementary with respect to quality. In this paper we propose a halftoning algorithm in black and white, derived from the error diffusion of Floyd Steinberg. By using a new threshold modulation, our new method combines the advantages of both masking and error diffusion algorithms. In order to evaluate our algorithm we defined a set of test images allowing the evaluation of the critical points of quality for the technical imagery: graininess, patterning and spatial bandwidth. The rendering of the presented algorithm has low graininess, no unpleasant patterning and broad spatial bandwidth. Tests were carried out with color images. We compared our algorithm with Floyd Stenberg's error diffusion with independent bitmaps plans. The improvement is similar to that we noticed in Black and white. 6058-38, Session 8 Graph order dither A. Hausner, Univ. of New Hampshire Printing technology produces the illusion of a continuous range of colors on devices that can produce only a few colors. This illusion can be achieved by dithering: arranging printed dots into patterns that the viewer's eye blends into the desired color. This paper presents a versatile generalization of dispersed-dot dithering that improves traditional dither methods, and also extends artistic screening techniques. Dithering has been studied intensively, but almost all past approaches assume the color dots are arranged in a square matrix. The present method works with arbitrarily-placed color points, and solves the following general problem: given an arbitrary set of 2D points (not necessarily on a grid) order them so that consecutive pairs and triples of points are as far apart as possible. The algorithm begins by building the adjacency graph of the points, and then obtains a graph ordering by applying graph coloring to its vertices recursively. Vertex coloring, a well-studied problem in graph theory, gives each graph vertex a color (or number) so that any two adjacent vertices will receive different colors. For every color, the subset of vertices with that color is considered. Each vertex in the subset is more separated from its nearest neighbor than nearest neighbors in the superset. The vertices in this subset are colored again, yielding points with even greater inter-neighbor distances; the process continues recursively. For each coloring, a color number is appended to each vertex's label. When recursion is done, each point has multi-digit label. Sorting these labels produces an ordering for the original points which has the desired consecutivepoint separation needed for dispersed dither. The above solution yields many useful results. Bayer's $2^n\times{}2^n$ dither matrices are obtained as a special case of the problem, when the initial points are on a square grid. However, dither matrices of arbitrary width and height can be also constructed. Moreover, since any given graph can be colored in many different ways, it is easy to obtain many different orderings for the same set of points. Using many orderings when printing an image avoids the obvious repetitive
artifacts evident in Bayer dither. In addition, reversing the digits in all the point labels and sorting can yield space-filling curves! Beyond improving classic dither, this method increases the versatility of existing artistic screening methods like Ostromoukhov's. His method repeats a single small motif image across the whole desired image, producing a delightful decorative effect that is also useful in counterfeit-resistant printing. His method is restricted to continuous-tone motifs, whereas the present method is more flexible, and works with any motif, including binary (black and white) images. To generate a dither pattern from a motif image, the present method considers each grey level occurring in the motif image. For each level the subset of pixels with that grey value is obtained. Each subset is treated as an arbitrary collection of 2D points and graphordered. Finally, all the ordered subsets are concatenated, yielding an ordering for all the motif's pixels for artistic dither. As a special case, applying graph ordering to a motif image whose brightness grows radially into its center will yield a dither matrix for clustereddot dither. Sample images are online at www.cs.unh.edu/~ah/hgc 6058-39, Session 8 Optimal halftoning over hexagonal grids J. Bacca Rodriguez, A. J. Gonzalez Lozano, G. R. Arce, Univ. of Delaware; D. L. Lau, Univ. of Kentucky The spectral analysis of blue noise dithering in hexagonal grids provides the desired spectral characteristics one must attain but it does not provide the dithering structures needed to achieve these. In this paper, these optimal dithering mechanisms are developed through modifications of the Direct Binary Search (DBS) algorithm extensively used for rectangular grids. Special attention is given to the effects of the new geometry on the Human Visual System (HVS) models and on the efficient implementation of the hexagonal-grid DBS. This algorithm provides the best possible output at the expense of high computational complexity, and while the DBS algorithm is not practical in most applications, it provides a performance benchmark for other more practical algorithms. Finally, a tone-dependent, hexagonal-grid, error-diffusion algorithm is developed, where the DBS algorithm is used to optimize the underlying filter weights. The characteristics of the HVS are thus implicitly used in the optimization. Extensive simulations show that hexagonal grids do indeed reduce disturbing artifacts, providing smoother halftone textures over the entire gray-scale region. Results also show that tone-dependent error-diffusion can provide comparable results to that of the DBS algorithms but at a significantly lower computational complexity.
112
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6059: Image Quality and System Performance III Tuesday-Thursday 17-19 January 2006 Part of Proceedings of SPIE Vol. 6059 Image Quality and System Performance III
6059-36, Poster Session Image quality assessment based on textual structure and noise normalization C. Zhang, Z. Qiu, Beijing Jiaotong Univ. (China) Considering the popular concept of objective image quality assessment models based on error sensitivity and Wang's assumption that HVS is highly adapted for extracting structural information, this paper proposed an objective image quality assessment SNPSNR based on textural structure and normalized noise. Taking time-frequency advantages of wavelet transform, SNPSNR describes the contribution of a coefficient by its energy proportion to corresponding approximation sub-band or detail subbands of a transform level, i.e., a frequency channel. The proportion is utilized to measure the distortion caused by the noise on that coefficient. HVS is also taken into consideration by weighting noises differently according to corresponding frequency channel. Due to the energy distribution property of wavelet transform, the noise quantity difference on each transform level is quite large and is not proportional to the influence caused by them. We normalize the structural noise on different levels by normalizing the coefficients on each level. Noise normalization on different frequency channels provides a consistent prediction of DMOS for images with different contents under different distortions. Finally the form of peak signalto-noise is used to compute SNPSNR. By comparisons with MSSIM, HVSNR and PSNR from fitting goodness with DMOS it turns out that SNPSNR models HVS better. 6059-37, Poster Session Quality models for audiovisual streaming T. C. Thang, Y. S. Kim, C. S. Kim, Y. M. Ro, Information and Communications Univ. (South Korea) Quality metric is an essential factor in multimedia communication, especially in compression and adaptation. Quality metric so far is mostly for a single modality (video, audio). There is a need to measure the quality of multimedia (i.e. multimodality) contents. Moreover, so far quality is just considered from the perceptual perspective. In practice, the content may be drastically adapted, even converted to another modality. In this case, we should consider the quality from semantic perspective as well. Previously, we show that multimedia quality has two aspects: perceptual quality and semantic quality. The former refers to user's satisfaction in perceiving the content, regardless of what information the content contains; the latter, which is crucial in modality conversion, refers to the amount of conveyed information, regardless of how the content is presented. Different applications may focus on different quality aspects (perceptual, semantic, or both). In this research, we consider the semantic quality of audiovisual contents; specifically, we estimate the amount of information conveyed by a videophone content where both video and audio channels may be strongly degraded, even audio is converted to text. We also consider the perceptual quality model of audio visual content, so as to see the difference with semantic quality model. To model the semantic quality of multimedia quality, we apply the concept of "conceptual graph", which consists of semantic nodes and relations between the nodes. Each node is attributed by a weight and the semantic quality of that node. And each relation is attributed by the strength of that relation. The semantic quality of an adapted content with respect to the original content is computed based on the similarity measure between the conceptual graphs of the adapted content and the original content. An audiovisual content can be modeled by a simple graph of two nodes (one for video and one for audio) and one relation between the nodes. The audiovisual semantic quality based on similarity
measure is essentially a parametric function (model) of the semantic quality of audio and the semantic quality of video. As shown in a previous work of ours, the semantic quality of a single modality (video, audio) can be modeled by some analytical function. Our focus here is how the semantic quality of audiovisual contents can be combined from the individual qualities of audio and video. To find the semantic quality model of audiovisual contents in videophone service, we design and carry out extensive subjective tests to measure semantic qualities of audio channel, video channel, and the audiovisual combination. The selected content type is headshoulder with speech, configured specifically for the practical videophone over wireless network. For an original audiovisual content, the video channel is adapted with four different frame-rates, while the audio channel is adapted with four different sampling rates and also converted to a textual transcription. Then 4 scores for video versions, 5 scores for audio versions, and 20 scores for audiovisual versions are recorded for each original audiovisual content. Finally, we obtain the specific quality model by multiple regression analysis with the subjective training data. Compared to a previous work on the perceptual quality model for audiovisual contents, the semantic quality model obtained in this work shows that the audio channel has very high impact on the overall quality of audiovisual quality. This can be explained by the fact that from the semantic perspective, the user is more interested in the audio channel. In addition, if we can quantify the semantic quality of textual transcription with respect to original audio channel, the obtained quality model can still be applied to measure the quality of adapted text-visual content. Further, the inter-subject reliability based on confidence intervals suggests that the obtained quality model can be reliably used in practical videophone service at low bandwidth. In the future, we will study the quality of other multimodality applications like e-learning, virtual reality. 6059-38, Poster Session Research on the analysis and measurement of MTF of staring imaging system D. Lu, Q. Chen, G. Gu, Nanjing Univ. of Science & Technology (China) The staring imaging technique is one of the main research directions in the field of the opto-electronic imaging. In order to design a thermal imaging system with good quality, a set of parameters are required for an objective evaluation of both static and dynamic performance of the imaging system. Static performance parameters are used to evaluate the static object imaging characteristics, namely, assuming that the object's spatial distribution does not vary with time (or the variation is negligible), whereas the dynamic ones describe the dynamic imaging capability. In general, following parameters are introduced to evaluate the thermal imaging system: MTF, NETD, MRTD, SNR, SiTF, uniformity, dynamic range, etc., among which MTF is one of the most important parameters in evaluating the performance of the detector and the system. In the MTF analysis, the object's spatial distribution is mapped to image's spatial distribution via convolution, which is one of the most important factors in the course of analyzing and evaluating the imaging performance. In this paper, we report a thorough analysis on the characteristics of MTF from the spatial and frequency spectrum. This Fourier transform based analysis was performed on the dynamic imaging characteristics of staring imaging system. In detail, two methods are introduced for MTF testing of the imaging system: the slit target and step target. Furthermore, abundant experiments are made on the measurement of MTF of visible CCD and IRFPA, thereby the results are obtained for the staring imaging system, and preliminary analysis is stated.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
113
Conf. 6059: Image Quality and System Performance III
6059-39, Poster Session Quantification method of the color breakup phenomena: evaluation of nextgeneration color wheels for field sequential color displays J. Thollot, K. Sarayeddine, Thomson R&D France (France); A. Trйmeau, Univ. Jean Monnet Saint-Etienne (France) While projection display market is growing up, micro-display based optical engines (e.g. DMD, LCD, LCOS) are continuously improved in order to make this technology competitive versus emergent large screen Liquid Crystal Displays and Plasma Displays . One of the main challenge is to make use of temporal integration capacity of Human Visual System by using only one single microdisplay. In such a case, the optical engine must include an illumination system that allows to time sequentially delivers a colored light flux which is then modulated by the micro-display. This illumination system can be based on a lamp to which a mechanical device (e.g. color wheel) is added in order to filter sequentially the emitted light flux. In any case, field sequential color displays exhibit a disturbing artifact known as the rainbow effect which is perceived as image content that breaks up into color components particularly next to high contrast edges. The present paper describes a psychophysical method developed by Thomson that quantifies the phenomena according to frame frequency, eye movement velocity and display technologies. Sample results are discussed for three and six primary field sequential color projection systems, showing that color ordering is a key to decrease the rainbow effect. 6059-40, Poster Session No-reference jerkiness evaluation method for multimedia communications M. Carli, D. Guida, A. Neri, Univ. degli Studi di Roma Tre (Italy) In this contribution, we present an objective assessment method for the evaluation, without reference, of the degradation of the video quality induced by the reduction of the temporal resolution. Jerkiness (or jerky motion) is a video coding artefact produced by temporal sub-sampling. In essence the motion of an object on the scene, that appears as smooth when played at the original frame rate, becomes stepwise with instantaneous abrupt changes when frame rate is drastically reduced. Subjective tests show that human observers are particularly sensitive to jerkiness. Several full reference techniques for jerkiness assessment have been presented in literature. The proposed assessment of the jerkiness perceived by a human observer is performed by feeding a multilayer neural network with the statistical distributions of the kinematics data (speed, acceleration and jerk of objects on the image plane) evaluated on a video shot. To identify the neural network (architecture and parameters) that best fit the human behaviour, a subjective experiment has been performed. Validation of the model on the test set indicate a good match between the Mean Opinion Score (MOS) and the jerkiness indicator computed by the neural network. 6059-01, Session 1 Fundamental questions related to print quality P. J. Mangin, M. Dubй, Univ. du Quйbec а Trois-Riviиres (Canada) The concept of print quality is elusive, since it depends on objective measurable quantities such as contrast, graininess, etc., but also on the subjective appreciation of potential observers. It is far from
obvious that print quality (PQ) can be defined in terms of good or bad, so that every one will agree on this definition, and the question of its measurement, with objective and subjective measures, remains open. In this Communication, we would first like to propose a set of fundamental questions related to the definition and measurement of PQ. Specifically, we are interested in the definition of PQ in terms of quality concept and quality criteria, on the minimal dimension space of PQ, and on the functional relations that should be satisfied by the metrics of PQ. In the second part, we focus on the simpler case of print mottle and try to answer some of these questions. We show that wavelet transforms can be used to obtain a measure of PQ that correlates very well with the subjective evaluation of observers and use this measure to discuss the functional form of a metric of Print Quality. 6059-02, Session 1 What do users really perceive: probing the subjective image quality G. S. Nyman, J. Radun, T. Leisti, Univ. of Helsinki (Finland); J. Oja, H. J. Ojanen, J. Olives, T. Vuori, Nokia Research Ctr. (Finland); J. P. Hдkkinen, Univ. of Helsinki (Finland) and Nokia Research Ctr. (Finland) Image evaluation schemes must satisfy both objective and subjective requirements. Objective image quality evaluation models are often preferred over subjective quality evaluation, because of their fastness and cost- effectiveness. However, the correlation between subjective and objective estimations is often poor. One of the key reasons for this is that it is not known what image features subjects use when they evaluate image quality. We have studied subjective image quality evaluation in the case of image sharpness. We used an Interpretation-Based Quality (IBQ) approach, which combines both qualitative and quantitative analysis of subjective data to probe the observer's quality experience. The combined data are obtained from the studies of subjective image sharpness, using five image contents (ISO 12640) and MTF manipulations of the test targets. We examine how naпve subjects experienced and classified natural images, whose sharpness was changing. Together the psychometric and qualitative information obtained allows the correlation of quantitative evaluation data with its underlying subjective attribute sets. This offers guidelines to product designers and developers who are responsible for image quality. Combining these methods makes the end- user experience approachable and offers new ways to improve objective image quality evaluation schemes. 6059-03, Session 2 The effect of image sharpness on quantitative eye-movement data and on image quality evaluation while viewing natural images T. Vuori, M. Olkkonen, Nokia Corp. (Finland) The aim of the study is to test both customer image quality rating (subjective image quality) and physical measurement of user behavior (eye movements tracking) to find customer satisfaction differences in imaging technologies. Methodological aim is to find out whether eye movements could be quantitatively used in image quality preference studies. In general, we want to map objective or physically measurable image quality to subjective evaluations and eye movement data. We conducted a series of image quality tests, in which the test subjects evaluated image quality while we recorded their eye movements. Results show that eye movement parameters consistently change according to the instructions given to the user, and according to physical image quality, e.g. saccade duration
114
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6059: Image Quality and System Performance III
increased with increasing blur. Results indicate that eye movement tracking could be used to differentiate image quality evaluation strategies that the users have. Results also show that eye movements would help mapping between technological and subjective image quality. Furthermore, these results give some empirical emphasis to top-down perception processes in image quality perception and evaluation by showing differences between perceptual processes in situations when cognitive task varies. 6059-04, Session 2 Assessing the enhancement of image sharpness S. Bouzit, Univ. of St. Andrews (United Kingdom); L. W. MacDonald, London College of Communication (United Kingdom) A psychophysical experiment was conducted to investigate the performance of four different image sharpness enhancement methods. Two were based on the power spectrum adjustment method using the human visual contrast sensitivity function and the other two methods applied PhotoShop's standard sharpening fitters. The results of the experiments are presented and discussed. Five major conclusions are drawn from this experiment: (1) Performance of sharpening methods; (2) Image dependence; (3) Influence of two different colour spaces on sharpness manipulation; (4) Correlation between perceived image sharpness and image preference; and (5) Effect of image sharpness enhancement on the image power spectrum. 6059-05, Session 2 Reference-free quality metric using a region-based attention model for JPEG2000 compressed images R. Barland, A. Saadane, Univ. de Nantes (France) At high compression ratios, JPEG-2000 coding introduces distortions (blurring, ringing) that potentially cause an embarrassment for a human observer. However, the Human Visual System does not carry out a systematic and local research of these impairments in the whole image, but rather, it identifies some regions of interest for judging the perceptual quality. In this paper, we propose to use both of these distortions (ringing and blurring effects), locally weighted by an importance map generated by a region-based attention model, to design a new reference free quality metric for JPEG-2000 compressed images. For the blurring effect, the impairment measure depends on spatial information contained in the whole image while, for the ringing effect, only the local information localized around strong edges is used. To predict the regions in the scene that potentially attract the human attention, a stage of the proposed metric consists to generate an importance map issued from the Osberger's attention model. The produced importance map is used to locally weight each distortion measure. The predicted scores have been compared on one hand, to the subjective scores and on other hand, to previous results, only based on the artefact measurements. This comparative study demonstrates the efficiency of the proposed quality metric. 6059-06, Session 2 Comparison of various subjective video quality assessment methods C. Lee, H. Choi, E. Lee, S. Lee, J. Choe, Yonsei Univ. (South Korea) As multimedia services, such as video on demand and video phones, over noisy channels are widely available to consumer market, quality monitoring emerges as an important topic. In this paper, we present comparison of three subjective testing methods: the double stimulus continuous quality scale (DSCQS) method, the
single stimulus continuous quality evaluation (SSCQE) method and the absolute category rating (ACR) method. The DSCQS method was used for validate objective models in the VQEG Phase II FR_TV test [1]. The SSCQE method is chosen to be used in the VQEG RRTV test [2]. The ACR method is chosen to be used in the VQEG Multimedia test [4]. Since a different subjective test method is used in each test, in-depth analyses of the three methods will provide helpful information in understanding human perception of video quality. In the double stimulus continuous quality scale method (DSCQS) recommended by ITU-R [3], evaluators are shown two video sequences: one is the original (source) video sequence (SRC) and the other is a processed video sequence (PVS). Without knowing which is the source video, the evaluators are shown the video sequences two times and is asked to provide their subjective scores by marking their subjecting rating in the provided form. Then, DMOS (Differential Mean Opinion Score) is computed by subtracting the score of the processed video sequence from that of the source video sequence. In the SSCQE method, a video sequence is shown only once to evaluators. The SSCQE method was conducted following the guideline of the VQEG RRTV test plan [2]. According to the test plan, video sequences may or may not contain impairments. Furthermore, there is a source video sequence for every PVS (processed video sequence) so that a hidden reference procedure can be implemented. In the SSCQE method, evaluators evaluate the video quality in real time using a slider which has a continuous grading scale. In accordance with the RRTV test plan [2], a test tape was created, which has a total of 60 video sequences. Each video sequence lasts one minute. The video sequences was generated by processing source video sequences (SRCs) using various HRCs (hypothetical reference circuits). Using test tape, subjective testing using the SSCQE method was conducted. Since source video sequences are always included, DMOS (Differential Mean Opinion Score) is also computed by subtracting the score of the processed video sequence from that of the source video sequence. In order to perform the DSCQS method, 64 8-second video sequences were taken from the 60-minutes test tape. Then, subjective test using the DSCQS was performed. These DSCQS scores were compared with the corresponding SSCQE scores, which were obtained by averaging the scores of the corresponding 8-second clip. It is noted that the SSCQE scores are MOS (mean opinion score). We can also compute SSCQE-DMOS by subtracting the score of the processed video sequence from that of the source video sequence. Using the 8-second video sequences, we also performed subjective testing using the ACR method. In the paper, we present statistical analyses of the three subjective testing methods. [1] VQEG, "Final Report from the Video Quality Experts Group on the Validation of Objective Models of Video Quality Assessment, Phase II (FR-TV2)," Aug 2003. [2] ITU-R Recommendation BT.500-10, Methodology for the subjective assessment of the quality of television pictures, 2000. [3] VQEG, "RRNR-TV Group TEST PLAN," www.vqeg.org, 2005. [4] VQEG, "Multimedia Group Test Plan," www.vqeg.org, 2005. 6059-07, Session 3 Selecting significant colors from a complex image for image quality modeling K. J. Leeming, P. J. Green, London College of Communication (United Kingdom) Judgements of complex images differ from those of uniform colour samples in several important respects. One such difference is that a complex image is formed of a large number of discrete colour elements. Observer judgements are based not on assessment of
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
115
Conf. 6059: Image Quality and System Performance III
each discrete element but of a much smaller number of salient features. The judgement process can be considered as the selection of such features followed by the judgement of particular quality attributes for these features. Modelling the judgement process thus requires a set of well-defined quality attributes together with a methodology for the selection of salient features and their relative importance. In this project, a method of selecting colours within an image was considered. A number of measurement locations within a complex image were selected, and the colour of these locations was measured on a series of reproductions. Measurements were carried out by registering each reproduction on an xy table and programming a 0:45 spectrophotometer with a 4mm aperture. The reproductions were judged by a panel of expert observers for their match to a proof, using a category scaling of several image quality attributes. By comparing the measured colour differences with the visual judgements it was possible to determine which locations carried the greatest weight in the assessments. It was also possible to make a limited prediction of the visual judgements from the measurements.By eliminating colour locations with the smallest weightings in predicting the visual judgements, the number of measurement locations was reduced to eight. Further analysis of the data looked at the tolerance to colour change when compared to the visual judgements for each of the selected areas. Caucasian skin tones were more tolerant to slight increases in red, having a central tendency of +1+ a*. Conversely the blue areas selected had a central tendency of -1+ a*. In a second phase of the project, further images were analysed to evaluate the ability of the salient colours identified in the first phase to predict the visual judgements of these images. In this phase, images were selected to have salient colours from phase 1 with both similar and different semantic content to phase 1. For both images it was determined that fewer colours were required to represent the image. It was deduced that this was due to both images being less complex than the image in phase 1. It was concluded that some image quality judgements can be predicted from a small number of salient colours, where the colours measured have similar semantic content. It was also found that the number of colours is to some extent image dependent. 6059-08, Session 3 Comparison of three full-reference color image quality measures E. Girshtel, V. Slobodyan, J. S. Weissman, A. M. Eskicioglu, The City Univ. of New York Image quality assessment plays a major role in many image processing applications. Although much effort has been made in recent years towards the development of quantitative measures, the relevant literature does not include many papers that have produced accomplished results. Ideally, a useful measure should be easy to compute, independent of viewing distance, and able to quantify all types of image distortions. In this paper, we will compare three fullreference full-color image quality measures (M-DFT, M-DWT, and MDCT). Assume the size of a given image is nxn. The transform (DFT, DWT, or DCT) is applied to the luminance layer of the original and degraded images. The transform coefficients are then divided into four bands, and the following operations are performed for each band: (a) obtain the magnitudes Moi, i=1,..., (nxn/4) of original transform coefficients, (b) obtain the magnitudes Mdi, i=1,..., (nxn/4) of degraded transform coefficients, (c) compute the absolute value of the differences: |Moi-Mdi|, i=1,..., (nxn/4), and (d) compute the standard deviation of the differences. Finally, the mean of the four standard deviations is obtained to produce a single value representing the overall quality of the degraded image. In our experiments, we have used five degradation types, and five degradation levels. The three proposed full-reference measures outperform the Peak-Signal-to-Noise Ratio (PSNR), and two stateof-the-art metrics Q and MSSIM.
6059-09, Session 3 Influence of ambient illumination on adapted and optimal white point I. Vogels, J. Berentsen, Philips Research Labs. (Netherlands) The white point recommended by the European Broadcasting Union (EBU) for TV displays, which corresponds to illuminant D65, does not always yield optimal color rendering. Previous research has shown that the most preferred white point strongly depends on image content. This study investigates the effect of the chromaticity and intensity of the ambient illumination on the adapted white point of a homogeneous image (i.e. the chromaticity that is perceived as achromatic) and on the optimal white point of natural images (i.e. the white point with the most preferred color rendering). It was found that the adapted white point and the optimal white point shift towards the chromaticity of the ambient illumination. The effect of illuminant color was approximately 2.5 times larger for the adapted white point than for the optimal white point. The intensity of the ambient illumination had no effect on the adapted white point and the optimal white point, except for images with face content. In agreement with previous studies, the optimal white point was found to depend on image content. The results indicate that the optimal color rendering of natural images is a complex relation of image content and ambient illumination. 6059-10, Session 4 Characterization of digital image noise properties based on RAW data H. H. Hytti, Tampereen Teknillinen Yliopisto (Finland) Noise properties of digital cameras are under intensive research all around the world. In current research project between TUT/MIT and Nokia corporation, image noise characterization based on digital camera RAW data is being studied. Digital image has several different noise sources. Separating these from each other, if possible, helps to improve image quality by reducing or even eliminating some noise components. It is impossible to completely separate the noise components from each other by analyzing RAW data, but by applying several different measurement and analysis procedures their nature and relative impact on image quality can be evaluated. In this paper, three different imaging technologies are compared. The three digital cameras used are Canon EOS D30 with CMOS sensor, Nikon D70 with CCD sensor and Sigma SD10 with Foveon X3 Pro 10M CMOS sensor. Due to different imaging sensor constructions, these cameras have rather different noise characteristics. The applicability of different analysis methods to these different sensor types is also studied. The analysis methods used in this research project include for example photon transfer curve method and ISO 15739 standard noise analysis methods. 6059-12, Session 4 An evaluation of sharpness in different image displays used for medical imaging M. Ukishima, T. Nakaguchi, Chiba Univ. (Japan); K. Kato, Canon Inc. (Japan); Y. Fukuchi, Chiba Univ. Hospital (Japan); N. Tsumura, Chiba Univ. (Japan); K. Matsumoto, Canon, Inc.; N. Yanagawa, H. Morita, Chiba Univ. Hospital (Japan); Y. Miyake, Chiba Univ. (Japan) X-ray film systems have been widely used for a diagnosis of various diseases since a long time ago. In recent years, many kinds of displays and recording systems for X-ray medical images have been used including inkjet printer, silver halide film, CRT and LCD, by the development of the digital X-ray image capturing systems. In this paper, image quality of X-ray images displayed onto high accurate
116
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6059: Image Quality and System Performance III
monochrome CRT and LCD monitors are analyzed and compared. Images recorded on the exclusive film and coated paper by inkjet printer and the wet type and dry type photo printers using a silver halide material are also analyzed and compared. The modified Gan's method is introduced to calculate the MTF (Modulation Transfer Function) from the knife ESF (edge spread function). The results show that the MTFs of the inkjet image on the transparency and the wet type silver halide film image have fairly similar and good response in comparison with the inkjet image on the coated paper and the dry type silver halide film. It is also shown that the CRT has the worse response over the spatial frequency range. It was well correlated between the MTF and observer rating value. From here, we consider the proposed method is effective. 6059-13, Session 5 Characterization of printer MTF W. Jang, Hewlett-Packard Co.; J. P. Allebach, Purdue Univ. We develop a comprehensive procedure for characterizing the modulation transfer function (MTF) of a digital printer. Especially designed test pages consisting of a series of patches, each with a different 1-D sinusoidal modulation, enable measurement of the dependence of the MTF on spatial frequency, bias point, modulation amplitude, spatial direction of modulation, and direction of modulation in the color space. Constant tone patches also yield the extreme and center color values for the input modulation. After calibrating the scanner specifically for the direction of modulation in the color space, we spatially project the scanned test patches in the direction orthogonal to the modulation to obtain a 1-D signal, and then project these sample points onto a line in the CIE L*a*b* color space between the extreme color values to obtain a perceptually relevant measure of the frequency response in a specific color direction. Appropriate normalization of the frequency response followed by compensation for the scanner MTF completes the procedure. For a specific inkjet printer using a dispersed-dot halftoning algorithm, we examine the impact of the abovementioned parameters on the printer MTF, and obtain results that are consistent with the expected behavior of this combination of print mechanism and halftoning algorithm. 6059-14, Session 5 PSF estimation by gradient descent fit to the ESF E. H. Barney Smith, Boise State Univ. Calibration of scanners and cameras usually involves measuring the point spread function (PSF). When edge data is used to measure the PSF, the differentiation step amplifies the noise. A parametric fit of the functional form of the edge spread function (ESF) directly to the measured edge data is proposed to eliminate this. Experiments used to test this method show that the Cauchy functional form fits better than the Gaussian or other forms tried. The effect of using a functional form of the PSF that differs from the true PSF is explored by considering bilevel images formed by thresholding. The amount of mismatch seen can be related to the difference between the respective kurtosis factors. 6059-15, Session 5 Printer banding estimation using the generalized spectrum N. A. Rawashdeh, I. Shin, K. D. Donohue, Univ. of Kentucky; S. T. Love, Lexmark International, Inc. This paper compares multi-step algorithms for estimating banding parameters of a harmonic signature model. The algorithms are based on two different spectral measures, the power spectrum (PS) and the collapsed average (CA) of the generalized spectrum. The
generalized spectrum has superior noise reduction properties and is applied for the first time to this application. Simulations compare estimation performances of profile (or coherent) averaging and noncoherent spatial averaging for estimating banding parameters in grain noise. Results demonstrate that profile averaging has superior noise reduction properties, but is less flexible in applications with irregular banding patterns. The PS-based methods result in lower fundamental frequency estimation error and greater peak height stability for low SNR values, with coherent averaging being significantly superior to non-coherent. The CA has the potential of simplifying the detection of multiple simultaneous banding patterns because its peaks are related to intra-harmonic distances; however, good CA estimation performance requires sufficiently regular harmonic phase patterns for the banding harmonics so as not to undergo reduction along with the noise. The algorithms are applied to samples from inkjet and laser printers to demonstrate the model's ability to separate banding from grain and other artifacts. 6059-16, Session 5 Scanner-based macroscopic color variation estimation C. Kuo, L. Di, E. K. Zeise, NexPress Solutions, Inc. Flatbed scanners have been adopted successfully in the measurement of microscopic image artifacts, such as granularity and mottle, in print samples because of their capability of providing full color, high resolution images. Accurate macroscopic color measurement relies on the use of colorimeters or spectrophotometers to provide a surrogate for human vision. The very different color response characteristics of flatbed scanners from any standard colorimetric response limits the utility of a flatbed scanner as a macroscopic color measuring device. This metamerism constraint can be significantly relaxed if our objective is mainly to quantify the color variations within a printed page or between pages where a small bias in measured colors can be tolerated as long as the color distributions relative to the individual mean values is similar. Two scenarios when converting color from the device RGB color space to a standardized color space such as CIELab are studied in this paper, blind and semi-blind color transformation, depending on the availability of the black channel information. We will show that both approaches offer satisfactory results in quantifying macroscopic color variation across pages while the semi-blind color transformation further provides fairly accurate color prediction capability. 6059-17, Session 6 Viewing conditions, colorimetric measurements, and profile making: the conundrum of standards vs. practical realities D. Q. McDowell, Standards Consultant The standards that define colorimetric measurements (ISO 13655) and viewing conditions (ISO 3664) for graphic arts and photography, and ICC profile building (ISO 15076-1) must all be consistent with each other. More importantly they must be consistent with current industry practice and be technically sound. However, as we begin the process of revising the color measurement and viewing standards we find that that is easier said than done. A black backing has traditionally been used for both measurement and viewing to minimize the effect of back printing and to be consistent with the black backing specified for densitometry. For profile building the color management community feels that measurements with white backing produce better results. The measurement community wants to compute both density and colorimetry from a single spectral reflectance measurement.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
117
Conf. 6059: Image Quality and System Performance III
Colorimetric measurements and viewing should match so the same backing needs to be used for viewing and measurement. Similarly the spectral power distribution of the illuminant needs to match between viewing and measurement. It is virtually impossible to make a measurement source that matches D50 in the UV. However, most papers used for proofing and printing use UV brighteners. Without UV these have no effect. Joint working groups have been formed by both ISO TC130 and ISO TC42 to find a solution to the conundrum that these issues present to the graphic arts, photographic and color management communities. 6059-18, Session 6 Progress in digital color workflow understanding in the ICC workflow WG A. L. McCarthy, Lexmark International, Inc. and International Color Consortium The ICC Workflow WG serves as a bridge between ICC color management technologies and use of those technologies in real world color reproduction applications. ICC color management is applicable to and is used in a wide range of color systems, from highly specialized digital cinema color effects to high volume publications printing to home use photography applications. The ICC Workflow WG works to align ICC technologies so that the color management needs of these diverse use case systems are addressed in an open, platform independent, manner. Over the past few years, the ICC Workflow WG has worked to formalize an understanding of color imaging and printing workflows. The focus of this presentation is a high level summary of the ICC Workflow WG objectives and work to date, as they impact image quality and color systems performance. Consider the questions, "How much of dissatisfaction with color management today is driven by `the wrong color transformation at the wrong time' or `I can't get to the right conversion at the right point in my work processes?" Put another way, consider how image quality through a workflow can be affected by the coordination and control level of the color management system. 6059-19, Session 6 Recent progress in the development of ISO 19751 S. Farnand, Eastman Kodak Co.; E. N. Dalal, Xerox Corp.; Y. S. Ng, NexPress Solutions, Inc. A small number of general visual attributes have been recognized as essential in describing image quality. These include micro-uniformity, macro-uniformity, colour rendition, text and line quality, gloss, sharpness, and spatial adjacency or temporal adjacency attributes. The multiple-part International Standard discussed here was initiated by the INCITS W1 committee on the standardization of office equipment to address the need for unambiguously documented procedures and methods, which are widely applicable over the multiple printing technologies employed in office applications, for the appearance-based evaluation of these visually significant image quality attributes of printed image quality. The resulting proposed International Standard, for which ISO/IEC WD 19751-1 presents an overview and an outline of the overall procedure and common methods, is based on a proposal that was predicated on the idea that image quality could be described by a small set of broad-based attributes. Five ad hoc teams were established (now six since a sharpness team is in the process of being formed) to generate standards for one or more of these image quality attributes. Updates on the colour rendition, text and line quality, and gloss attributes are provided.
6059-20, Session 6 ISO 19751 macro-uniformity R. S. Rasmussen, Xerox Corp.; K. D. Donohue, Univ. of Kentucky; Y. S. Ng, NexPress Solutions, Inc.; W. C. Kress, Toshiba America DSE; S. Zoltner, Xerox Corp.; F. Gaykema, OCE Technologies BV (Netherlands) The ISO WD 19751 macro-uniformity team works towards the development of a standard for evaluation of perceptual image quality of color printers. The team specifically addresses the types of defects that fall in the category of macro-uniformity, such as streaks, bands and mottle. The first phase of the standard will establish a visual quality ruler for macro-uniformity, using images with simulated macro-uniformity defects. A set of distinct, parameterized defects has been defined, as well as a method of combining the defects into a single image. The quality ruler will be a set of prints with increasing magnitude of the defect pattern. The paper will discuss the creation and printing of the simulated images, as well as initial tests of subjective evaluations using the ruler. 6059-21, Session 7 Edge-raggedness evaluation using slantededge analysis P. D. Burns, Eastman Kodak Co. The standard ISO 12233 method for the measurement of spatial frequency response (SFR) for digital still cameras and scanners is based on the analysis of slanted-edge image features. The procedure, which applies a form edge-gradient analysis to an estimated edge spread function, requires the automated finding of an edge feature in a digital test image. A frequently considered (e.g., ISO 13660 and 19751) attribute of printed text and graphics is edge raggedness. There are various metrics aimed at the evaluation of the discontinuous imaging of nominally continuous features, but they generally rely on an estimation of the spatial deviation of edge or line boundaries, the tangential edge profile (TEP). In this paper, we describe how slanted-edge analysis can be adapted to the routine evaluation of line and edge quality. After locating and analyzing the edge feature, the TEP is estimated. The estimation of RMS deviation and edge spectrum are described. 6059-22, Session 7 Statistical interpretation of ISO TC42 dynamic range: risky business D. R. Williams, P. D. Burns, Eastman Kodak Co. Recently, two ISO electronic imaging standards aimed at digital capture device dynamic range metrology have been issued. Both ISO 15739 (digital still camera noise) and ISO 21550 (film scanner dynamic range) adopt a signal-to-noise ratio (SNR) criterion for specifying dynamic range. The exposure levels that correspond to threshold-SNR values are used as endpoints to determine measured dynamic range. While these thresholds were developed through committee consensus with generic device applications in mind, the methodology of these standards is flexible enough to accommodate different application requirements. This can be done by setting the SNR thresholds according to particular signal-detection requirements. We provide an interpretation of dynamic range that can be related to particular applications based on contributing influences of variance, confidence intervals, and sample size variables. In particular, we introduce the role of the spatialcorrelation statistics for both signal and noise sources, not covered in previous discussions of these ISO standards. It is this frequency aspect to dynamic range evaluation that may well influence future standards. This is important when comparing systems with different sampling settings, since the above noise statistics are currently computed on a per-pixel basis.
118
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6059: Image Quality and System Performance III
6059-25, Session 8 The influence of statistical variations on image quality B. O. Hultgren, D. W. Hertel, Consultant; J. Bullitt, Polaroid Corp. For more than thirty years imaging scientists have constructed metrics to predict psychovisually perceived image quality. Such metrics are based on a set of objectively measurable basis functions such as Noise Power Spectrum (NPS), Modulation Transfer Function (MTF), and characteristic curves of tone and color reproduction. Although these basis functions constitute a set of primitives that fully describe an imaging system from the standpoint of information theory, we found that in practical imaging systems the basis functions themselves are determined by system-specific primitives, i.e. technology parameters. In the example of a printer MTF and NPS are largely determined by dot structure. In addition MTF is determined by color registration, and NPS by streaking and banding. Since any given imaging system is only a single representation of a class of more or less identical systems, the family of imaging systems and the single system are not described by a unique set of image primitives. For an image produced by a given imaging system, the set of image primitives describing that particular image will be a singular instantiation of the underlying statistical distribution of that primitive. If we know precisely the set of imaging primitives that describe the given image we should be able to predict its image quality. Since only the distributions are known, we can only predict the distribution in image quality for a given image as produced by the larger class of `identical systems' We will demonstrate the combinatorial effect of the underlying statistical variations in the image primitives on the objectively measured image quality of a population of printers as well as on the perceived image quality of a set of test images. We also will discuss the choice of test image sets and impact of scene content on the distribution of perceived image quality. 6059-26, Session 8 The use of a virtual printer model for the simulation of imaging systems B. O. Hultgren, Consultant A companion paper discusses the impact of statistical variability on perceived image quality. Early in a development program, systems may not be capable of rendering images suitable for quality testing. This does not diminish the program need to estimate the perceived quality of the imaging system. During the development of imaging systems, simulations are extremely effective for demonstrating the visual impact of design choices, allowing both the development process to prioritize these choices and management to understand the risks and benefits of such choices. Where the simulation mirrors the mechanisms of image formation, it not only improves the simulation but also informs the understanding of the image formation process. Clearly the simulation process requires display or printing devices whose quality does not limit the simulation. We will present a generalized methodology. When used with common profile making and color management tools, it will provide simulations of both source and destination devices. The device to be simulated is modeled by its response to a fixed set of input stimuli. In the case of a digital still camera (DSC), these are the reflection spectra of a fixed set of color patches -e.g. the MacBeth DCC, and in the case of a printer, the set of image RGBs. We will demonstrate this methodology with examples of various print media systems.
6059-27, Session 8 Improved pen alignment for bidirectional printing E. Bernal, J. P. Allebach, Z. Pizlo, Purdue Univ. The advent of low-cost, photo-quality inkjet printers has raised the need for an objective means of determining print quality that is consistent with what the end-user perceives. The ultimate objective of automated quality assessment processes is to provide the ability to measure a large volume of prints and at the same time, achieve the repeatability and objectivity that visual inspection-based processes lack. While some literature discusses metrics for the objective evaluation of print quality, few of the efforts have combined automated quality tests with subjective assessment. We develop an algorithm for analyzing printed dots and study the effects of the dot characteristics on the perceived print alignment. We demonstrate, via a set of psychophysical experiments, that the human viewer judges alignment of two line segments based on the position at which the edges of the segments reach a certain level of absorptance, rather than on the position of the segment centroids. We also show that the human viewer is less sensitive to changes in alignment as the irregularity of the dots that compose the line segments grows. 6059-28, Session 9 Further image quality assessment in digital film restoration M. Chambah, Univ. de Reims Champagne-Ardenne (France); C. Saint Jean, Univ. de La Rochelle (France); F. Helt, Independent Consultant (France) Several digital film restoration techniques have emerged during the last decade and became more and more automated but restoration evaluation still remains a rarely tackled issue. In the sphere of cinema, the image quality is judged visually. In fact, experts and technicians judge and determine the quality of the film images during the calibration (post production) process. As a consequence, the quality of a movie is also estimated subjectively by experts in the field of digital film restoration. On the other hand, objective quality metrics do not necessarily correlate well with perceived quality. Plus, some measures assume that there exists a reference in the form of an "original" to compare to, which prevents their usage in digital restoration field, where often there is no reference to compare to. That is why subjective evaluation is the most used and most efficient approach up to now. But subjective assessment is expensive, time consuming and does not respond, hence, to the economic requirements. After presenting the several defects than can affect cinematographic material, and the film digital restoration field, we present in this paper the issues of image quality evaluation in the field of digital film restoration and suggest some reference free objective measures. 6059-29, Session 9 Development of picture quality monitoring system for IPTV service based on the reduced reference framework O. Sugimoto, R. Kawada, A. Koike, KDDI R&D Labs. (Japan) With the rapid spread of high speed Internet access lines such as ADSL and FTTH, various novel network applications have been presented recently. IPTV (Internet Protocol based Television) service is one of these novel network applications. IPTV service has the advantage of offering lower cost broadcasting service via a high speed and low-cost Internet access line; however, it also has a
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
119
Conf. 6059: Image Quality and System Performance III
disadvantage in that picture quality degradation is caused by quality degradation of the transmission line such as IP packet loss. From the standpoint of application service providers, it is important to constantly monitor picture quality degradation to comprehend user level QoS and thus the demand for technologies and systems that enable automatic picture quality measurement has been increasing recently. The Reduced Reference (RR) method is especially effective among the currently available frameworks for picture quality assessment. The RR method is a framework for objective picture quality measurement that extracts image features from the baseband signal of the transmitted and received pictures and transmits them to a remote monitoring facility to compare the transmitted image features. This framework is being endorsed by ITU study groups and is currently under study for standardization. We previously studied a method to estimate the objective picture quality of PSNR based on the RR framework. The proposed method utilizes spread spectrum and orthogonal transform to extract image features and enables PSNR estimation with 0.984 correlation coefficient just by extracting image features at a rate of 320bit/frame (corresponding to a data channel's bitrate of 9.6kbps) from transmitted and received pictures. This shows the proposed method has sufficient estimation accuracy to monitor the picture quality of IPTV services. In this paper, we attempt to prove the effectiveness of the proposed method by developing a system based on the previous study and monitoring objective picture quality of the commercial IPTV service. As the contents delivery network (CDN) for the target IPTV service consists of connections of certain numbers of ring networks and the broadcasting contents are transmitted by IP Multicast, the picture quality on each receiver (user) side is different; it depends on the quality of the network route, i.e. packet loss rate at the users' terminal. Therefore, the proposed system monitors frame-by-frame PSNR between the sender side and a number of receivers simultaneously and the network operators are alerted when a transmission failure occurs. One of the advantages of the proposed system is that the proposed system reduces computational complexity by the method shown in the previous study and can be entirely implemented by PC software. This means the system can run in real time without any additional hardware whereas most of the existent picture quality assessment methods do not consider real-time operation. In this paper, we first describe the method to estimate PSNR based on the previous study and then describe the configuration of the proposed system. Finally, the results of the experiment, which continuously monitors the PSNR of the transmitted pictures at four different receiver points over 48 hours, are presented to demonstrate the effectiveness of the proposed system. 6059-30, Session 9 An image quality evaluation method for DOE-based camera lenses S. Lee, Y. Jin, H. Jeong, Samsung Electro-Mechanics Co., Ltd. (South Korea); S. Song, Hanyang Univ. (South Korea) A novel image quality evaluation method based on the rigorous grating diffraction theory and the ray-optic method is proposed and is applied for design optimization and tolerance analysis of DOEbased optical imaging system. First, diffraction efficiencies for various diffracted orders were calculated for RGB wavelengths. Secondly, a virtual resolution chart was generated to form an object and the output image was obtained by the CODE V(tm). The final image was obtained by summation of all images obtained above weighted by their diffraction efficiencies. Effects of fabrication errors such as the profile thickness and the shoulder on image quality are shown to be effectively predicted by this method. A DOE-based 2Mresolution phone-camera lens module shows ~15% MTF
improvement compared with a design through CODE V(tm) without such an optimization. Additional analysis shows ~12% degradation in MTF for the 150 lp/mm with a 6 micron shoulder length for the lens module. 6059-32, Session 10 Visibility and annoyance of LCD defective subpixels of different colors and surrounds at different positions H. Ho, J. M. Foley, S. K. Mitra, Univ. of California/Santa Barbara In this study we investigate the visibility and annoyance of simulated defective sub-pixels in a liquid crystal display (LCD). We carried out a psychophysical experiment to examine the effects of four variables: surround luminance, surround type, defect sub-pixel color and defect-pixel position. The stimulus was a rectangular image containing one centered object with a gray surround and a single defective pixel. The surround was either uniform gray or a gray-level texture. The target was a simulated discolored pixel with one defective sub-pixel (green, red or blue) and two normally functioning sub-pixels. On each trial, it was presented at a random position. Subjects were asked to indicate if they saw a defective pixel, and if so, where it was located and how annoying it was. For uniform surrounds, our results show that detection probability falls slowly for green, faster for red, and fastest for blue as background luminance increases. When detection probability is plotted against luminance contrast green defective pixels are still most detectable, then red, then blue. Mean annoyance value falls faster than detection probability as background luminance increases, but the trends are the same. A textured surround greatly reduces the detection probability of all defective pixels. Still, green and red are more detectable than blue. With the textured surround the mean annoyance tends to remain high even when detection probability is quite low. For both types of surrounds, probability of detection is least for targets in the bottom region of the image. 6059-33, Session 10 Robust detection of defects in imaging arrays J. Dudas, C. G. Jung, G. H. Chapman, Simon Fraser Univ. (Canada); Z. Koren, I. Koren, Univ. of Massachusetts/Amherst As digital imagers continue to increase in size and pixel density, the detection of faults in the field becomes critical to delivering high quality output. Traditional schemes for defect detection utilize specialized hardware at the time of manufacture and are impractical for use in the field, while previously proposed software-based approaches tend to lead to quality-degrading false positive diagnoses. This paper presents an algorithm that utilizes statistical information extracted from a sequence of normally captured images to identify the location and type of defective pixels. Building on previous research, this algorithm utilizes data local to each pixel and Bayesian statistics to more accurately infer the likelihood of each defect, which successfully improves the detection time. Several defect types are considered, including pixels with one-half of the typical sensitivity and permanently stuck pixels. Monte Carlo simulations have shown that for defect densities of up to 0.5%, 50 ordinary images are sufficient to accurately identify all faults without falsely diagnosing good pixels as faulty. Testing also indicates that the algorithm can be extended to higher resolution imagers and to those with noisy stuck pixels, with only minimal cost to performance.
120
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6059: Image Quality and System Performance III 6059-34, Session 10 Objective video quality assessment method for freeze distortion based on freeze aggregation K. Watanabe, J. Okamoto, T. Kurita, NTT Service Integration Labs. (Japan) With the development of the broadband network, video communications such as videophone, video distribution, and IPTV services are beginning to become common. In order to provide these services appropriately, we must manage them based on subjective video quality, in addition to designing a network system based on it. Currently, subjective quality assessment is the main method used to quantify video quality. However, it is timeconsuming and expensive. Therefore, we need an objective quality assessment technology that can estimate video quality from video characteristics effectively. Video degradation can be categorized into two types: spatial and temporal. Objective quality assessment methods for spatial degradation have been studied extensively, but methods for temporal have hardly been examined even though it occurs frequently due to network degradation and has a large impact on subjective quality. In this paper, we propose an objective quality assessment method for temporal. Our approach is to aggregate multiple freeze distortions into an equivalent freeze distortion and then derive the objective video quality from the equivalent freeze distortion. We also propose a method using the perceptual characteristics of short freeze distortions. We verified that our method can estimate the objective video quality well within the deviation of subjective video quality. 6059-35, Session 10 Film grain synthesis and its applications for re-graining P. Schallauer, R. Mцrzinger, JOANNEUM RESEARCH GmbH (Austria) Digital film restoration and special effects compositing require more and more automatic procedures for movie re-graining. Missing or inhomogeneous grain decreases perceived quality. For the purpose of grain synthesis an existing texture synthesis algorithm has been evaluated and optimized. We show that this algorithm can produce synthetic grain which is perceptually similar to a given grain template, which has high spatial and temporal variation and which can be applied to multi-spectral images. Furthermore a re-grain application framework is proposed, which synthesises based on an input grain template artificial grain and composites this together with the original image content. Due to its modular approach this framework supports manual as well as automatic re-graining applications. Two example applications are presented, one for regraining an entire movie and one for fully automatic re-graining of image regions produced by restoration algorithms. Low computational cost of the proposed algorithms allows application in industrial grade software.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
121
Conf. 6060: Visualization and Data Analysis 2006 Monday-Tuesday 16-17 January 2006 Part of Proceedings of SPIE Vol. 6060 Visualization and Data Analysis 2006
6060-01, Session 1 Multiscale image based flow visualization A. C. Telea, Technische Univ. Eindhoven (Netherlands); R. Strzodka, Research Ctr. Caesar (Germany) We present MIBFV, a method to produce real-time, multiscale animations of flow datasets. MIBFV extends the attractive features of the Image-Based Flow Visualization (IBFV) method, i.e. dense flow domain coverage with flow-aligned noise, real-time animation, implementation simplicity, and few (or no) user input requirements, to a multiscale dimension. We generate a multiscale of flow-aligned patterns using an algebraic multigrid method and use them to synthesize the noise textures required by IBFV. We demonstrate our approach with animations that combine multiple scale noise layers, in a global or level-of-detail manner. 6060-02, Session 1 Visualizing oceanic and atmospheric flows with streamline splatting Y. Sun, E. Ess, D. Sapirstein, M. Huber, Purdue Univ. The investigation of the climate system is one of the most exciting areas of scientific research today. In the climate system, oceanic and atmospheric flows play a critical role. Because these flows are very complex in the span of temporal and spatial scales, effective computer visualization techniques are crucial to the analysis and understanding of the flows. However, the existing techniques and software are not sufficient to the demand of visualizing oceanic and atmospheric flows. In this paper, we use a new technique called streamline splatting to visualize 3D flows. This technique integrates streamline generation with the splatting method of volume rendering. It first generates segments of streamlines and then projects and splats the streamline segments onto the image plane. The projected streamline segments can be represented using a Hermite parametric model. Splatted curves are achieved by applying a Gaussian footprint function to the projected streamline segments and the results are blended together. Thus the user can see through a volumetric flow field and obtain a 3D representation view in one image. This work has potential to be further developed into visualization software for regular PC workstations to help researchers explore and analyze climate flows. 6060-03, Session 1 View-dependent multiresolutional flow texture advection L. Li, H. Shen, The Ohio State Univ. Existing texture advection techniques will produce unsatisfactory rendering results when there is a discrepancy between the resolution of the flow field and that of the output image. This is because many existing texture advection techniques such as Line Integral Convolution (LIC) are inherently none view-dependent, that is, the resolution of the output textures depends only on the resolution of the input field, but not the resolution of the output image. When the resolution of the flow field after projection is much higher than the screen resolution, aliasing will happen unless the flow textures are appropriately filtered through some expensive post processing. On the other hand, when the resolution of the flow field is much lower than the screen resolution, a blocky or blurred appearance will be present in the rendering because the flow texture does not have enough samples. In this paper we present a view-dependent multiresolutional flow texture advection method for structured rectiand curvi-linear meshes. Our algorithm is based on a novel intermediate representation of the flow field, called trace slice, which allows us to compute the flow texture at a desired resolution
interactively based on the run-time viewing parameters. As the user zooms in and out of the field, the resolution of the resulting flow texture will adapt automatically so that enough flow details will be presented while aliasing is avoided. Our implementation utilizes mipmapping and programmable GPUs available on modern programmable graphics hardware. 6060-04, Session 2 Volumetric depth peeling for medical image display D. M. Borland, J. P. Clarke, J. R. Fielding, R. M. Taylor II, The Univ. of North Carolina at Chapel Hill Volumetric depth peeling (VDP) is an extension to volume rendering that enables display of otherwise occluded features in volume data sets. VDP decouples occlusion calculation from the volume rendering transfer function, enabling independent optimization of settings for rendering and occlusion. The algorithm is flexible enough to handle multiple regions occluding the object of interest, as well as object self-occlusion, and requires no pre-segmentation of the data set. VDP was developed as an improvement for virtual arthroscopy for the diagnosis of joint trauma, and has been generalized for use in non-invasive urology studies. In virtual arthroscopy, the surfaces in the joints often occlude each other, allowing limited viewpoints from which to evaluate these surfaces. In urology studies, the physician would like to position the virtual camera outside the kidney collecting system and see inside it. By rendering invisible all voxels between the observer's point of view and objects of interest, VDP enables viewing from unconstrained positions. Radiologists using VDP display have been able to perform evaluations of pathologies more easily and more rapidly than with clinical arthroscopy, standard volume rendering, or standard MRI/CT slice viewing. 6060-05, Session 2 Adaptive border sampling for hardware texture-based volume visualization E. C. LaMar, Institute for Scientific Research, Inc. We introduce a technique to properly sample volume boundaries in hardware texture-based Volume Visualization. Prior techniques render a volume with a set of uniformly-spaced proxy geometries that sample (and represent) a set of uniform-depth slices. While this is sufficient for the core of a volume, it does not consider a sample's *partial* overlap at the boundaries of a volume; and this failing can lead to significant artifactsat these boundaries. Our technique computes the non-unit depth contributions of the volume at the boundaries. We use fragment programs to compute the partial sample contributions and to match sampling-planes at the volume boundaries to the sampling geometry in the core of the volume. 6060-06, Session 2 Ray-casting time-varying volume data sets with frame-to-frame coherence D. Tost, S. Grau, Univ. Politиcnica de Catalunya (Spain); M. Ferre, Univ. Rovira i Virgili (Spain); A. Puig, Univ. de Barcelona (Spain) The goal of this paper is the proposal and evaluation of a raycasting strategy for time-varying volume data, that takes advantage of the spatial and temporal coherence in image-space as well as in object-space in order to speed up rendering. The proposed algorithm is based on a double structure: in image-space, a temporal buffer that stores for each pixel the next instant of time in which the pixel must be recomputed and, in object-space, a Temporal Run-Length Encoding of the voxel values through time.
122
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6060: Visualization and Data Analysis 2006
The algorithm skips empty and unchanged pixels through three different space-leaping strategies. It can compute the images sequentially in time or generate them simultaneously in batch. In addition, it can handle simultaneously several data modalities. Finally, an on-purpose out-of-core strategy is used to handle large datasets. The tests performed on two medical datasets and various phantom datasets show that the proposed strategy significantly speeds-up rendering. 6060-07, Session 3 Theoretical analysis of uncertainty visualizations T. D. Zuk, M. S. T. Carpendale, Univ. of Calgary (Canada) Although a number of theories and principles have been developed to guide the creation of visualizations, it is not always apparent how to apply the knowledge in these principles. We describe the application of perceptual and cognitive theories for the analysis of uncertainty visualizations. General principles from Bertin, Tufte, and Ware are outlined and then applied to the analysis of eight different uncertainty visualizations. The theories provided a useful framework for analysis of the methods, and provided insights into the strengths and weaknesses of various aspects of the visualizations. 6060-08, Session 3 A visualization framework for design and evaluation B. J. Blundell, S. Pettifer, The Univ. of Manchester (United Kingdom); G. Ng, Cerebra, Inc. The creation of compelling visualisation paradigms is a craft often dominated by intuition and issues of aesthetics, with relatively few models to support good design. The majority of problem cases are approached by simply applying a previously evaluated visualisation technique. A large body of work exists covering the individual aspects of visualisation design such as the human cognition aspects visualisation methods for specific problem areas, psychology studies and so forth, yet most frameworks regarding visualisation are applied after-the-fact as an evaluation measure. We present an extensible framework for visualisation aimed at structuring the design process and increase decision traceability, delineating the notions of function, aesthetics and usability. The framework can be used to deriving a set of requirements for good visualisation design, and evaluating existing visualisations, presenting possible improvements. Our framework achieves this by being both broad and general, built on top of existing works, with hooks for extensions and customizations. This paper shows how existing theories of information visualisation fit into the scheme, share our experience in the application of this framework on several designs, and offering our evaluation of the framework and the designs studied. 6060-09, Session 4 Visual analytics and the NVAC P. C. Wong, Pacific Northwest National Lab. No abstract available 6060-10, Session 5 Maximum entropy lighting for physical objects T. Malzbender, E. Ordentlich, Hewlett-Packard Labs. This paper presents a principled method for choosing informative lighting directions for physical objects. An ensemble of images of an
object or scene is captured, each with a known, predetermined lighting direction. Diffuse reflection functions are then estimated for each pixel across such an ensemble. Once these are estimated, the object or scene can be interactively relit as it would appear illuminated from an arbitrary lighting direction. We present two approaches for evaluating images as a function of lighting direction. The first uses image compressibility evaluated across a grid of samples in lighting space. Two compression techniques are evaluated, the first being Huffman encoding, the second being JPEG-LS. Highly compressible images using either algorithm have low information content, and lighting directions associated with these images fail to be informative. The second approach for choosing lighting directions uses image variance and prediction error variance, which are monotonically related to compressibility for Gaussian distributions. The advantage of the variance approach is that both image variance and prediction error variance can be analytically derived from the scene reflection functions, and evaluated at the rate of a few nanoseconds per lighting direction. 6060-11, Session 5 Pre-computed illumination for isosurfaces K. M. Beason, Florida State Univ.; J. Grant, Pixar Animation Studios; D. C. Banks, B. Futch, M. Y. Hussaini, Florida State Univ. Commercial software systems are available for displaying isosurfaces (also known as level sets, implicit surfaces, varieties, membranes, or contours) of 3D scalar-valued data at interactive rates, allowing a user to browse the data by adjusting the isovalue. We present a technique for applying global illumination to the resulting scene by pre-computing the illumination for level sets and storing it in a 3D illumination grid. The technique permits globally illuminated surfaces to be rendered at interactive rates on an ordinary desktop computer with a 3D graphics card. We demonstrate the technique on datasets from magnetic resonance imaging (MRI) of the human brain, confocal laser microscopy of neural tissue in the mouse hippocampus, computer simulation of a Lennard-Jones fluid, and computer simulation of a neutron star. 6060-12, Session 5 Retro-rendering with vector-valued light: producing local illumination from the transport equation D. C. Banks, K. M. Beason, Florida State Univ. We demonstrate that local illumination (without shadows or interreflections) can be interpreted as the correct solution to the equation for light transport by offering a novel interpretation of luminaire emittance and surface reflectance. 6060-13, Session 6 Bit-plane based analysis of integer wavelet coefficients for image compression A. F. Abu-Hajar, Digitavid, Inc. This paper presents bit-plane based statistical study for integer wavelet transforms commonly used in image compression. In each bit-plane, the coefficients were modeled as binary random variables. Experimental results indicate the probability of the significant coefficients (P1), in each bit-plane, monotonically increases from P1 0 at the most significant bits (MSB) to p1 0.5 at the least significant bits (LSB). Then, a parameterized model to predict p1 from the MSB to the LSB was proposed. Also, the correlation among the different bit-planes within the same coefficient was investigated. In addition, this study showed correlation of the significant coefficients in the same spatial orientation among different subbands. Finally, clustering within the each subband and across the different subband with the same spatial orientation was
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
123
Conf. 6060: Visualization and Data Analysis 2006
investigated. Our results show strong correlation of previously coded significant coefficients at higher levels and the significant coefficients in future passes at lower levels. The overall study of this paper is useful in understanding and enhancing existing waveletbased image compression algorithms such as SPIHT and EBC.
6060-16, Session 7 Real-time 3D visualization of DEM combined with a robust DCT-based datahiding method
6060-14, Session 6 Two-dimensional reduction PCA: a novel approach for feature extraction, representation, and recognition R. M. Mutelo, W. L. Woo, S. S. Dlay, Univ. of Newcastle Upon Tyne (United Kingdom) Recently, an image based technique coined two-dimensional principal component analysis (2DPCA) was developed for image representation and recognition. Although 2DPCA is superior to PCA in terms of computational efficiency for feature extraction and recognition accuracy, 2DPCA-based image representation is not as efficient as PCA in terms of storage requirements, since 2DPCA requires more coefficients for image representation than the feature vector for PCA. Our work shows that the feature matrix for 2DPCA has very little noise and a redundancy between its columns and the relevant data is compressed within a few columns. Therefore the noise and redundancy between the feature matrix rows still exist leading not only a large storage requirement but also to poor classification accuracy and large classification time. Therefore, we design a sequential optimal compression mechanism which eliminating redundancies and noise in two directions, rows and columns. Thus resulting in computational efficient, greater recognition accuracy, more efficient memory storage and lastly reduced classification time. The superiority of our algorithm is further demonstrated by the experimental results conducted on the ORL face database where an accuracy of 95.0% using a 9 by 5 for the proposed and 93.0% using 112 by 7 for 2DPCA. 6060-15, Session 7 Energetically optimal travel across terrain: visualizations and a new metric of geographic distance with archaeological applications B. Wood, Harvard Univ.; Z. Wood, California Polytechnic State Univ. We present a visualization and computation tool for modeling the caloric cost of pedestrian travel across three dimensional terrains. This tool is being used in ongoing archaeological research that analyzes how costs of locomotion affect the spatial distribution of trails and artifacts across archaeological landscapes. Throughout human history, traveling by foot has been the most common form of transportation, and therefore analyses of pedestrian travel costs are important for understanding prehistoric patterns of resource acquisition, migration, trade, and political interaction. Traditionally, archaeologists have measured geographic proximity based on "as the crow flies" distance. We propose new methods for terrain visualization and analysis based on measuring paths of least caloric expense, calculated using well established metabolic equations. Our approach provides a human centered metric of geographic closeness, and overcomes significant limitations of available Geographic Information System (GIS) software. We demonstrate such path computations and visualizations applied to archaeological research questions. Our system includes tools to visualize: energetic cost surfaces, comparisons of the elevation profiles of shortest paths versus least cost paths, and the display of paths of least caloric effort on Digital Elevation Models (DEMs). These analysis tools can be applied to calculate and visualize 1) likely locations of prehistoric trails and 2) expected ratios of raw material types to be recovered at archaeological sites.
A. Martin, Univ. Montpellier II (France); G. Gesquiere, Univ. de Provence (France); W. Puech, Univ. Montpellier II (France); S. Thon, Univ. de Provence (France) Using aerial photography, satellite imagery, scanned maps and Digital Elevation Models implies to make storage and visualization strategy choices. To obtain a three dimensional visualization, we have to link these images called texture with the terrain geometry named Digital Elevation Model. These information are usually stored in three different files (One for the DEM, one for the texture and one for the geo-referenced coordinates). In this paper we propose to store these information in only one file. In order to solve this problem, we present a technique for color data hiding of images, based on DC components of the DCT-coefficients. In our application the images are the texture, and the elevation data are hidden in each block. This method mainly protects against JPEG compression and cropping. 6060-17, Session 8 Hierarchical causality explorer: making complemental use of 3D/2D visualizations S. Azuma, Ochanomizu Univ. (Japan); I. Fujishiro, Tohoku Univ. (Japan); H. Horii, The Univ. of Tokyo (Japan) Hierarchical causality relationships reside ubiquitously in the reality. Since the relationships take intricate forms with two kinds of links -- hierarchical abstraction and causal association, there exists no single visualization style that allows the user to comprehend them effectively. This paper introduces a novel information visualization framework which can change existing 3D and 2D display styles interactively according to the user's visual analysis demands. The two visualization styles play a complementary role, and the change in the style relies on morphing so as to maintain the user's cognitive map. Based on this framework, we have developed a generalpurpose prototype system, which provides the user with an enriched set of functions not only for supporting fundamental information seeking, but bridging analytic gaps to accomplishing high-level analytic tasks such as knowledge discovery and decision making. The effectiveness of the system is illustrated with an application to the analysis of a nuclear-hazard cover-up problem. 6060-18, Session 8 InvIncrements: incremental software to support visual simulation D. C. Banks, W. Blanco, Florida State Univ. We describe a set of incremental software modules, based on Open Inventor, that support an interdisciplinary course in interactive 3D simulation. 6060-19, Session 9 Plot of plots and selection glass H. Chen, SAS Institute Inc. Modern dynamic data visualization environments often feature complex displays comprised of many interactive components, such as plots, axes, and others. These components typically contain attributes or properties that can be manipulated programmatically or
124
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6060: Visualization and Data Analysis 2006
interactively. Component property manipulation is usually a twostage process. The user first selects or in some way identifies the component to be revised and then invokes some other technique or procedure to modify the property of interest. Until recently, components typically have been manipulated one at a time, even if the same property is being modified in each component. How to effectively select multiple components interactively in multiple-view displays remains an open issue. This paper proposes modeling the display components with conventional data sets and reusing simple dynamic graphics, such as a scatter plot or a bar chart, as the graphical user interface to select these elements. This simple approach, called plot of plots, provides a uniform, flexible, and powerful scheme to select multiple display components. In addition, another approach called selection glass is also presented. The selection glass is a tool glass with click-on and click-through selection tool widgets for the selection of components. The availability of the plot of plots and selection glass provides a starting point to investigate new techniques to simultaneously modify the same properties on multiple components. 6060-20, Session 9 Navigation techniques for large-scale astronomical exploration C. Fu, The Hong Kong Univ. of Science and Technology (Hong Kong China); A. J. Hanson, E. A. Wernert, Indiana Univ. Navigating effectively in virtual environments at human scales is a difficult problem. However, it is even more difficult to navigate in large-scale virtual environments such as those simulating the physical Universe; the huge spatial range of astronomical simulations and the dominance of empty space make it hard for users to acquire reliable spatial knowledge of astronomical contexts. This paper introduces a careful combination of navigation and visualization techniques to resolve the unique problems of largescale real-time exploration in terms of travel and wayfinding. For large-scale travel, spatial scaling techniques and constrained navigation manifold methods are adapted to the large spatial scales of the virtual Universe. We facilitate large-scale wayfinding and context awareness using visual cues such as power-of-10 reference cubes, continuous exponential zooming into points of interest, and a scalable world-in-miniature (WIM) map. These methods enable more effective exploration and assist with accurate context-model building, thus leading to improved understanding of virtual worlds in the context of large-scale astronomy. 6060-21, Session 10 Reducing InfoVis cluttering through non uniform sampling, displacement, and user perception E. Bertini, G. Santucci, L. Dell'Aquila, Univ. degli Studi di Roma/La Sapienza (Italy) Clutter affects almost any kind of visual technique and can obscure the structure present in the data even in small datasets, making it hard for users to find patterns and reveal relationships. In this paper we present a general strategy to analyze and reduce clutter using a special kind of sampling, together with an ad-hoc displacement technique and perceptual issues collected through a user study. The method, defined for 2D scatter plots, is flexible enough to be used in quite different contexts. In particular, in this paper we prove its usefulness against scatter plot, radviz, and parallel coordinates visualizations.
6060-22, Session 10 Diverse information integration and visualization S. L. Havre, A. R. Shah, C. Posse, B. M. Webb-Robertson, Pacific Northwest National Lab. This paper presents and explores a technique for visually integrating and exploring diverse information. Researchers and analysts seeking knowledge and understanding of complex systems have increasing access to related, but diverse, data. These data provide an opportunity to simultaneously analyze entities of interest from multiple informational perspectives through the integration of diverse but related data. Our approach visualizes an entity set across multiple perspectives; each is an alternate partitioning of the entity set. The partitioning may be based on inherent or assigned attributes such as meta-data or prior knowledge. The partitioning may also be derived directly from entity data, for example, through unsupervised classification, or clustering. The same entities may be clustered on data from different experiment types or processing approaches. This reduction of diverse data/information on an entity to a series of partitions, or discrete (and unit-less) categories, allows the user to view the entities across diverse data without concern for data types and units. Parallel coordinate plots typically visualize continuous data across multiple dimensions. We adapt parallel coordinate plots for discrete values such as partition names to allow comparison of entity patterns across multiple dimensions to identify trends and outlier entities. We illustrate this approach through a prototype. 6060-23, Session 10 WordSpace: visual summary of text corpora U. Brandes, M. Hoefer, J. Lerner, Univ. of Konstanz (Germany) In recent years several well-known approaches to visualize the topical structure of a document collection have been proposed. Most of them feature spectral analysis of a term-document matrix with influence values and dimensionality reduction. We generalize this approach by arguing that there are many reasonable ways to project the term-document matrix into low-dimensional space in which different features of the corpus are emphasized. Our main tool is a continuous generalization of adjacency-respecting partitions called structural similarity. In this way we obtain a generic framework in which influence weights in the term-document matrix, dimensionality-reducing projections, and the display of a target subspace may be varied according to nature of the text corpus. 6060-24, Session 11 Information architecture: why design matters J. Agutter, Univ. of Utah No abstract available 6060-25, Session 12 Trees in a treemap: visualizing multiple hierarchies M. Burch, S. Diehl, Katholische Univ. Eichstдtt (Germany) This paper deals with the visual representation of a particular kind of structured data: trees where each node is associated with an object (leaf node) of a taxonomy. We introduce a new visualization technique that we call Trees in a Treemap. In this visualization edges can either be drawn as straight or orthogonal edges. We compare our technique with several known techniques. To demonstrate the
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
125
Conf. 6060: Visualization and Data Analysis 2006
usability of our visualization techniques, we apply them to two kinds of real world data: software project data and network routing data. We show both patterns and anomalies which can be found by using our visualization. To improve the readability of the graphical representation we use different color codings for nodes and edges. Furthermore we try to minimize the number of edge crossings, the number of edge direction changes and the length of the edges, three contrary goals. 6060-26, Session 12 Focus-based filtering + clustering technique for power-law networks with small world phenomenon F. Boutin, J. Thiиvre, M. Hascoлt, Univ. Montpellier II (France) and CNRS (France) Realistic interaction networks usually present two main properties: power-law degree distribution and small world behaviour. Few nodes are linked to many nodes and adjacent nodes are likely to share common neighbours. Moreover, graph structure usually presents a dense core that is difficult to explore with classical filtering and clustering techniques. In this paper, we propose a new filtering technique that takes into account a user-focus. This technique extracts a tree-like graph that also has power-law degree distribution and small world behaviour. Resulting structure is easily drawn with classical force-directed drawing algorithms. It is also quickly organised and displayed into a nested tree of clusters from any userfocus. We built a new graph filtering + clustering + drawing API and report a case study. 6060-27, Session 12 Enhancing scatterplot matrices for data with ordering or spatial attributes Q. Cui, M. O. Ward, E. A. Rundensteiner, Worcester Polytechnic Institute The scatterplot matrix is one of the most common methods used to project multivariate data onto two dimensions for display. While each off-diagonal plot maps a pair of non-identical dimensions, there is no prescribed mapping for the diagonal plots. In this paper, histograms, 1D plots and 2D plots are drawn in the diagonal plots of the scatterplots matrix. In 1D plots, the data are assumed to have order, and they are projected in this order. In 2D plots, the data are assumed to have spatial information, and they are projected onto locations based on these spatial attributes using color to represent the dimension value. The plots and the scatterplots are linked together by brushing. Brushing on these alternate visualizations will affect the selected data in the regular scatterplots, and vice versa. Users can also navigate to other visualizations, such as parallel coordinates and glyphs, which are also linked with the scatterplot matrix by brushing. Ordering and spatial attributes can also be used as methods of indexing and organizing data. Users can select an ordering span or a spatial region by interacting with 1D plots or with 2D plots, and then observe the characteristics of the selected data subset. 1D plots and 2D plots provide the ability to explore the ordering and spatial attributes, while other views are for viewing the abstract data. In a sense, we are linking what are traditionally seen as scientific visualization methods with methods from the information visualization and statistical graphics fields. We validate the usefulness of this integration by providing two case studies, time series data analysis and spatial data analysis.
6060-28, Session 13 Content-based text mapping using multidimensional projections for exploration of document collections R. Minghim, F. V. Paulovich, A. de Andrade Lopes, Univ. de Sгo Paulo (Brazil) This paper presents a technique for generation of maps of documents targeted at placing similar documents in the same neighborhood. As a result, besides being able to group (and separate) documents by their contents, it runs at very manageable computational costs. Based on multi-dimensional projection techniques and an algorithm for projection improvement, it results in a surface map that allows the user to identify a number of important relationships between documents and sub-groups of documents via visualization and interaction. Visual attributes such as height, color, isolines and glyphs as well as aural attributes (such as pitch), help add dimensions for integrated visual analysis. Exploration and narrowing of focus can be performed using a set of tools provided. This novel text mapping technique, named IDMAP (Interactive Document Map), is fully described in this paper. Results are compared with dimensionality reduction and cluster techniques for the same purposes. The maps are bound to support a large number of applications that rely on retrieval and examination of document collections and to complement the type of information offered by current knowledge domain visualizations. 6060-29, Session 13 Mapping texts through dimensionality reduction and visualization techniques for interactive exploration of document collections A. d. A. de Andrade Lopes, R. Minghim, V. Melo, F. V. Paulovich, Univ. de Sгo Paulo (Brazil) This paper presents a methodology to create a meaningful graphical representation of documents corpora targeted at supporting exploration of correlated documents. The purpose of such an approach is to produce a map from a document body on a research topic or field based on the analysis of their contents, and similarities amongst articles. The document map is generated, after text preprocessing, by projecting the data in two dimensions using Latent Semantic Indexing. The projection is followed by hierarchical clustering to support sub-area identification. The map can be interactively explored, helping to narrow down the search for relevant articles. Tests were performed using a collection of documents pre-classified into three research subject classes: CaseBased Reasoning, Information Retrieval, and Inductive Logic Programming. The map produced was capable of separating the main areas and approaching documents by their similarity, revealing possible topics, and identifying boundaries between them. The tool can deal with the exploration of inter-topics and intra-topic relationship and is useful in many contexts that need deciding on relevant articles to read, such as scientific research, education, and training. 6060-30, Session 14 Visualizing brain rhythms and synchrony K. A. Robbins, D. Veljkovic, E. Pilipaviciute, The Univ. of Texas at San Antonio Patterns of synchronized brain activity have been widely observed in EEGs and multi-electrode recordings, and much study has been devoted to understanding their role in brain function. We introduce
126
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6060: Visualization and Data Analysis 2006
the problem of visualization of synchronized behavior and propose visualization techniques for assessing temporal and spatial patterns of synchronization from data. We discuss spike rate plots, activity succession diagrams, space-time activity band visualization, and low-dimensional projections as methods for identifying synchronized behavior in populations of neurons and for detecting the possibly short-lived neuronal assemblies that produced them. We use wavelets conjunction with these visualization techniques to extract the frequency and temporal localization of synchronized behavior. Most of these techniques can be streamed, making them suitable for analyzing long-running experimental recordings as well as the output of simulation models. The visualizations shown is this paper have been implemented in DAVIS (Data Viewing System: http:// visual.cs.utsa.edu/research/Davis/index.html), visualization software that supports a large number of simultaneous synchronized views of data with different sampling rates. The results of this paper show the usefulness of visualization in studying synchronization of neuronal populations on the order of a thousand neurons. Wavelet-based spatial multiscale techniques may also be useful in tackling larger populations. 6060-31, Session 14 Automatic feature-based surface mapping for brain cortices L. Linsen, Ernst Moritz Arndt Univ. Greifswald (Germany) We present a method that maps a complex surface geometry to an equally complicated, similar surface. One main objective of our effort is to develop technology for automatically transferring surface annotations from an atlas brain to a subject brain. While macroscopic regions of brain surfaces often correspond, the detailed surface geometry of corresponding areas can vary greatly. We have developed a method that simplifies a subject brain's surface forming an abstract yet spatially descriptive point cloud representation, which we can match to the abstract point cloud representation of the atlas brain using an approach that iteratively improves the correspondence of points. The generation of the point cloud from the original surface is based on surface smoothing, surface simplification, surface classification with respect to curvature estimates, and clustering of uniformly classified regions. Segment mapping is based on spatial partitioning, principal component analysis, rigid affine transformation, and warping based on the thin-plate spline (TPS) method. The result is a mapping between topological components of the input surfaces allowing for transfer of annotations. 6060-32, Poster Session Blogviz: mapping the dynamics of information diffusion in blogspace M. S. Lima, Parsons School of Design Blogviz is a visualization model for mapping the transmission and internal structure of top links across the blogosphere. It explores the idea of meme propagation by assuming a parallel with the spreading of most cited URLs in daily weblog entries. The main goal of Blogviz is to unravel hidden patterns in the topics diffusion process. What's the life cycle of a topic? How does it start and how does it evolve through time? Are topics constrained to a specific community of users? Who are the most influential and innovative blogs in any topic? Are there any relationships amongst topic proliferators?
6060-33, Poster Session Organizing and visualizing database data using parallel coordinates C. G. Presser, Gettysburg College In this paper, we describe a data organization and axis grouping technique for managing parallel coordinate plots. A database visualization model is created as an intermediary between the data and the visualization. On the visualization side, axes within a parallel coordinate plot are put into groups which can be represented by a new axis in the plot, while the members of the group are hidden. Methods are presented for building these groups and displaying their axes, each with their own advantages and disadvantages. Lastly, a working system which uses these techniques to visualize data from a database is presented. 6060-35, Poster Session Visualizing 3D vector fields with splatted streamlines E. Ess, Y. Sun, Purdue Univ. We present a novel technique called streamline splatting to visualize 3D vector fields interactively. This technique integrates streamline generation with the splatting method of volume rendering. The key idea is to create volumetric streamlines using geometric streamlines and a kernel footprint function. To optimize the rendering speed, we represent the volumetric streamlines in terms of a series of slices perpendicular to the principal viewing direction. Thus 3D volume rendering is achieved by blending all slice textures with support of graphics hardware. This approach allows the user to visualize 3D vector fields interactively such as by rotation and zooming on regular PCs. This new technique may lead to better understanding of complex structures in 3D vector fields. 6060-36, Poster Session SRS browser: a visual interface to the sequence retrieval system K. K. Mane, K. Borner, Indiana Univ. This paper presents a novel approach to the visual exploration and navigation of complex association networks of biological data sets, e.g., published papers, gene or protein information. The generic approach was implemented in the SRS Browser as an alternative visual interface to the highly used Sequence Retrieval System (SRS) [1]. SRS supports keyword-based search of about 400 biomedical databases. While the SRS presents search results as rank-ordered lists of matching entities, the SRS Browser displays entities and their relations for interactive exploration. A formal usability study was conducted to examine the SRS Browser interface's capabilities to support knowledge discovery and management. 6060-39, Poster Session Tracing parallel vectors J. Sukharev, Univ. of California/Santa Cruz Feature tracking algorithms usually rely on operators for identifying regions of interest. One of these commonly used operators is to identify parallel vectors introduced by Peikert and Roth In this paper, we propose a new and improved method for finding parallel vectors in 3D vector fields. Our method uses a two-stage approach where in the first stage we extract solution points from 2D faces using Newton-Raphson method, and in the second stage, we use analytical tangents to trace solution lines. The distinct advantage of our method over the previous method lies in the fact that our
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
127
Conf. 6060: Visualization and Data Analysis 2006
algorithm does not require a very fine grid to find all the important topological features. As a consequence, the extraction phase does not have to be at the same resolution as the original dataset. More importantly, the feature lines extracted are topologically consistent. We demonstrate the tracing algorithm with results from several datasets. 6060-40, Poster Session Output-sensitive volume tracking L. Jiang, Rutgers Univ. Feature tracking is a useful technique for studying the evolution of phenomena (or features) in time-varying scientific datasets. Timevarying datasets can be massive and are constantly becoming larger as more powerful machines are being used for scientific computations. To interactively explore such datasets, feature tracking must be done efficiently. For massive datasets, which do not fit into memory, tracking should be done out-of-core. In this paper, we propose an "output-sensitive" feature tracking, which uses the pre-computed metadata to (1) enable out-of-core processing structured datasets, (2) expedite the feature tracking processing, and (3) make the feature tracking less threshold sensitive. With the assistance of the pre-computed metadata, the complexity of the feature extraction is improved from O(mlgm) to O(n), where m is the number of cells in a timestep and n is the number of cells in just the extracted features. Furthermore, the feature tracking's complexity is improved from O(nlgn) to O(nlgk), where k is the number of cells in a feature group. The metadata computation and feature tracking can easily be adapted to the outof-core paradigm. The effectiveness and efficiency of this algorithm is demonstrated using experiments. 6060-41, Poster Session Visualization of force fields in protein structure prediction S. N. Crivelli, Lawrence Berkeley National Lab. and California Institute for Quantitative Biomedical Research; C. Crawford, O. Kreylos, B. Hamann, Univ. of California/Davis One of the challenges of protein structure prediction and folding is the uncertainty of modeling the energetic forces causing the folding. This is particularly important because the native structure of a protein corresponds to the global minimum of its energy function. There is a need to develop interactive visualization tools that allow researchers to study, compare, and minimize energy functions. Unfortunately, these energy functions are mathematically defined in ways that do not facilitate a straightforward application of visualization techniques. To visualize energy, it is necessary to define a spatial mapping for these force fields. Such a mapping would allow generating volume renderings of the internal energy states of a molecule. We describe the spatial mapping that we use for energy, and the visualizations that we produce from this mapping. We integrated energy visualization into ProteinShop, an interactive graphical environment for the manipulation of protein structures that are used as initial configurations for further minimization. This enables us to manipulate protein configurations guided by energy and to observe changes in gradient during minimization. We will provide images and animations offering insight into the biological behavior of the proteins as well as the computational behavior of the energy optimization algorithms we use.
6060-42, Poster Session Correspondence-based visualization techniques M. J. Gerald-Yamasaki, NASA Ames Research Ctr. A visual representation model is an abstract pattern used to create images which characterize quantitative information. By using a texture image to define a visual representation model, correspondence of color to denote similarity, and correspondence of image location over multiple images to associate information into collections, highly effective visualization techniques are made possible. One such technique for two-dimensional texture-based vector field visualization is vector field marquetry. Vector field marquetry uses a synthesized image representing direction as a conditioner for pixel replacement over a collection of vector field direction-magnitude portraits. The resulting synthesized image displays easily recognizable local and global features, vector direction, and magnitude. A related technique enabled by correspondence-based methods is the sparse representation of a vector field by a topological skeleton constructed from isodirection lines. Each vector in a vector field along an isodirection line points in the same direction. Isodirection lines subdivide the domain into regions of similar vectors, converge at critical points, and represent global characteristics of the vector field.
128
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6061: Internet Imaging VII Wednesday-Thursday 18-19 January 2006 Part of Proceedings of SPIE Vol. 6061 Internet Imaging VII
6061-30, Poster Session Subjective trajectory characterization: acquisition, matching, and retrieval M. Y. Zhang, L. Olsen, J. E. Boyd, Univ. of Calgary We describe a system that automatically tracks moving objects in a scene and subjectively characterize an object trajectory for storage and retrieval. A multi-target color-histogram particle filter combined with data association is the foundation of our trajectory acquisition algorithm and Procrustes shape analysis is the basis of our subjective trajectory representation. Particle filters are useful for tracking the state of non-linear dynamic systems with clutter and noise. The adaptive color-histogram particle filter uses color measurements to track objects under changes in view and lighting in complex scenes. However, tracking multiple targets requires multiple particle filters, and the automatic initialization and termination of the individual filters presents a challenge. We handle initialization and termination by coupling with simple motion-based object detection. e.g., background subtraction. The system attempts to associate each detected object with a particle filter state. When an object appears that does not associate with a particle filter, the system initializes a new filter. When a particle filter does not have associated objects for an extended time, the system terminates the filter. One can view this as a variation of a multiple-hypothesis tracker that: (a) uses particle filters instead of Kalman filters, and (b) uses zero scan-back (i.e., resolves the data association at each time step). To improve computational performance, we use quasi-Monte-Carlo methods to reduce the number of particles required by each filter. The tracking system operates in real-time to produce a stream of XML documents that contain the object trajectories. To characterize trajectories subjectively, we form a set of shape templates that describes basic maneuvers (e.g., gentle turn right, hard turn left, straight line). Procrustes shape analysis provides a scale- and rotation-invariant mechanism to identify occurrences of these maneuvers within a trajectory. Thus, a temporal sequence of basic maneuvers describes the shape of a trajectory. To add spatial information to our trajectory representation, we partition the twodimensional space under surveillance into a set of mutually exclusive regions. A temporal sequence of region-to-region transitions gives a spatial representation of the trajectory. The shape and position descriptions combine to form a compact, highlevel representation of a trajectory. We provide similarity measures for the shape, position, and combined shape and position representations. Finally, we present experimental results that demonstrate the advantages of this approach for indexing and retrieval in a trajectory database. 6061-32, Poster Session Archiving of meaningful scenes for personal TV terminals S. H. Jin, J. H. Cho, Y. M. Ro, Information and Communications Univ. (South Korea); J. Kim, Electronics and Telecommunications Research Institute (South Korea) In this paper, we propose an archiving method of broadcasts for TV terminals including a set-top box (STB) and a personal video recorder (PVR). Our goal is to effectively cluster and retrieve semantic video scenes obtained by real-time broadcasting content filtering for re-use or transmission. For TV terminals, we generate new video archiving formats which combine broadcasting media resources with the related metadata and auxiliary media data. In addition, we implement an archiving system to decode and retrieve the media resource and the metadata within the format. The
experiment shows that the proposed format makes it possible to retrieve or browse media data or metadata in the TV terminal effectively, and could have compatibility with a portable device. 6061-33, Poster Session AVIR: a spoken document retrieval system in e-learning environment I. Gagliardi, M. Padula, P. Pagliarulo, Consiglio Nazionale delle Ricerche (Italy) The development of automatic methods for Spoken Document Retrieval (SDR) continues to emerge as an important research area for both the speech and information retrieval communities. In this paper we present AVIR (Audio & Video Information Retrieval), a project of CNR (Italian National Research Council) - ITC to develop a tools to support an information system for distance e-learning. AVIR has been designed for the storage, indexing, classification and retrieval of audio and video lessons and teaching materials to make them available to students and other interested users. The core of AVIR is a SDR (Spoken Document Retrieval) system which automatically transcribes the spoken documents into texts and indexes them through dictionaries appropriately created (taxonomies, ontologies), so that the users can retrieve the material of interest by means of textual queries. During the fruition on-line, the user can formulate his queries searching documents by date, professor, title of the lesson or selecting one or more specific words. The results are presented to the users: in case of video lessons the preview of the first frames is shown. Moreover, slides of the lessons and associate papers can be retrieved. 6061-34, Poster Session Internet-based remote counseling to support stress management: preventing interruptions to regular exercise in elderly people S. Hashimoto, Univ. of Tsukuba (Japan) and National Institute of Information and Communications Technology (Japan); T. Munakata, Univ. of Tsukuba (Japan); N. Hashimoto, Citizen Watch Co., Ltd. (Japan); J. Okunaka, T. Koga, National Institute of Information and Communications Technology (Japan) We used an internet based, remotely conducted, face to face, preventive counseling program using video monitors to reduce the source of life-stresses that interrupts regular exercise and evaluated the preventative effects of the program. NTSC Video signals were converted to the IP protocol and facial images were transmitted to a PC display using the exclusive optical network lines of JGN2. Participants were 22 elderly people. IT remote counseling was conducted on two occasions. A survey was conducted before the intervention in August 2003, a post survey in February 2004 and a follow-up in March 2005. Network quality was satisfactory with little data loss and high display quality. Results indicated that self-esteem increased significantly, trait anxiety decreased significantly, cognition of emotional support by people other than family members had a tendency to increase, and source of stress had a tendency to decrease after the intervention. Follow-up results indicated that cognition of emotional support by family increased significantly, and interpersonal dependency decreased significantly compared to before the intervention. These results suggest that IT remote counseling is useful to keep elderly people from feeling anxious and to make them confident to continue exercising regularly. Moreover, it has a stress management effect.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
129
Conf. 6061: Internet Imaging VII
6061-35, Poster Session Vertex and face permutation order compression for efficient animation support E. Chang, Electronics and Telecommunications Research Institute (South Korea); D. Kim, B. Min, S. Lee, Hanyang Univ. (South Korea); N. Hur, S. Lee, Electronics and Telecommunications Research Institute (South Korea); E. S. Jang, Hanyang Univ. (South Korea) In this paper, we have proposed a new vertex and face permutation order compression algorithm to address the efficient animation support with the minimal size of side information. The main contributions of this paper are two: 1) provision of vertex and face permutation order, 2) compression of vertex and face permutation order. Our proposed vertex and face permutation order coding algorithm is based on both the adaptive probability model instead of static probability model and the connected component representation to achieve coding efficiency. As a result, we can allocate one-less-bit codeword to each vertex and face permutation order in every distinguishable unit as encoding process proceeds. And, representing and encoding the given vertex and face permutation order in the unit of the connected component can lead to the reduction of required bit-rate. Test results demonstrate that the proposed algorithm can encode the vertex and face permutation order losslessly while making up to 12% bit-saving compared to the logarithmic representation based on the fixed probability. Furthermore, experimental results demonstrate that further improvement using the connected component concept brings 38% bit-saving compared to the proposed method and 46% bit-saving compared to the logarithmic encoding. 6061-01, Session 1 Requirements for benchmarking personal image retrieval systems J. Bouguet, C. Dulong, I. V. Kozintsev, Intel Corp. It is now common to have accumulated tens of thousands of personal pictures. Efficient access to that many pictures can only be done with a robust image retrieval system. This application is of high interest to processor architects. It is highly compute intensive, and could motivate end users to upgrade their personal computers to the next generations of processors. A key question is how to assess the robustness of a personal image retrieval system. Personal image databases are very different from digital libraries that have been used by many Content Based Image Retrieval Systems. Personal image databases are very different from digital libraries that have been used by many Content Based Image Retrieval Systems. For example a personal image database has a lot of pictures of people, but a small set of different people typically family, relatives, and friends. Pictures are taken in a limited set of places like home, work, school, and vacation destination. The most frequent queries are searched for people, and for places. These attributes, and many others affect how a personal image retrieval system should be benchmarked, and benchmarks need to be different from existing ones based on art images, or medical images for examples. The attributes of the data set do not change the list of components needed for the benchmarking of such systems as specified in: data sets, query tasks, ground truth, and evaluation measures. This paper proposes a way to build these components to be representative of personal image databases, and of the corresponding usage models.
6061-02, Session 1 On usage models of content-based image search, filtering, and annotation D. Telleen-Lawton, C. B. Chang, VIMA Technologies, Inc.; E. Y. Chang, Univ. of California/Santa Barbara VIMA has observed increasing end-user demand for Content-based Image Retrieval (CBIR) systems since late 2004. This paper's objective is to provide to image retrieval researchers and developers a framework for selecting the best performance measurements for their systems and algorithms based on users' applications and requirements. We describe CBIR search, filtering, and annotation systems, outline their applications and process flow, and provide details of their usage models. We also enumerate some technical challenges of CBIR systems and outline solutions for some of these challenges. 6061-03, Session 2 Human factors in automatic image retrieval system design and evaluation A. Jaimes, Fuji Xerox Co., Ltd. (Japan) Image retrieval is a human-centered task: images are created by people and are ultimately accessed and used by people for humanrelated activities. In designing image retrieval systems and algorithms, or measuring their performance, it is therefore imperative to consider the conditions that surround both the indexing of image content and the retrieval. This includes examining the different levels of interpretation for retrieval, possible search strategies, and image uses. Furthermore, we must consider different levels of similarity and the role of human factors such as culture, memory, and personal context. This paper takes a human-centered perspective in outlining levels of description, types of users, search strategies, image uses, and human factors that affect the construction and evaluation of automatic content-based retrieval systems, such as human memory, context, and subjectivity. 6061-04, Session 2 Lessons from TRECVID: lexicon design for semantic indexing in media databases M. R. Naphade, IBM Thomas J. Watson Research Ctr. No abstract available 6061-21, Session 2 Benchmarking without ground truth S. Santini, Univ. of California/San Diego Many evaluation techniques for content based image retrieval are based on the availability of a ground truth, that is on a "correct" categorization of images so that, say, if the query image is of category A, only the returned images in category A will be considered as "hits." Based on such a ground truth, standard information retrieval measures such as precision and recall and given and used to evaluate and compare retrieval algorithms. Coherently, the assemblers of benchmarking data bases go to a certain length to have their images categorized. The assumption of the existence of a ground truth is, in many respect, naive. It is well known that the categorization of the images depends on the a priori (from the point of view of such categorization) subdivision of the semantic field in which the images are placed (a trivial observation: a plant subdivision for a botanist is very different from that for a layperson). Even within a given semantic field, however, categorization by human subjects is subject to uncertainty, and it
130
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6061: Internet Imaging VII
makes little statistical sense to consider the categorization given by one person as the unassailable ground truth. (Even worse if, as it often happens, the person who creates the categorization is involved in the design of the system that is being tested!). In this paper we propose several evaluation techniques that apply to the case in which the ground truth is subject to uncertainty. In this case, obviously, measures such as precision and recall as well will be subject to uncertainty. The paper will explore the relation between the uncertainty in the ground truth and that in the most commonly used evaluation measures, so that the measurements done on a given system can preserve statistical significance. The work will also touch briefly on the influence of such uncertainty on the creation and organization of benchmarking data bases. 6061-06, Session 3 Using heterogeneous annotation and visual information for the benchmarking of image retrieval system H. Mьller, Univ. Hospital of Geneva (Switzerland) Many image retrieval systems and evaluation methodologies make use of either visual or textual information. Only few combine text and visual features for retrieval and evaluation. If text is used, it often relies upon standardised and complete annotation schema for the entire collection. This, in combination with high-level semantic queries, makes visual/textual combinations almost useless as the information need can often be solved using just text. In reality, many collections have some form of annotation but often heterogeneous and incomplete. Web-based image repositories such as FlickR even allow collective, as well as multilingual annotation. This article describes the ImageCLEF evaluation campaign. Unlike some other evaluations, we offer a range of realistic tasks and image collections in which combining text and visual features is likely to obtain the best results. We also offer a medical task which models the situation of heterogenous annotation by combining four collections with annotations of varying quality, structure, extent and language. Two collections have an annotation per case, which is normal in the medical domain, making it difficult to relate parts of the accompanying text to an image. This is typical of image retrieval from the web in which adjacent text does not always describe an image. ImageCLEF shows the need for realistic and standardised datasets, search tasks and ground truths for image retrieval evaluation. 6061-07, Session 3 On benchmarking content-based image retrieval applications B. Zhang, Y. Zuo, Tsinghua Univ. (China) No abstract available 6061-08, Session 3 TRECVID: the utility of a content-based video retrieval evaluation A. G. Hauptmann, Carnegie Mellon Univ. TRECVID, an annual retrieval evaluation benchmark organized by NIST, encourages research in information retrieval from digital video. TRECVID benchmarking covers both interactive and manual searching by end users, as well as the benchmarking of some supporting technologies including shot boundary detection, extraction of semantic features, and the automatic segmentation of TV news broadcasts. Evaluations done in the context of the TRECVID benchmarks show that generally, speech transcripts and annotations provide the single most important clue for successful
retrieval. However, automatically finding the individual images is still a tremendous and unsolved challenge. The evaluations repeatedly found that none of the multimedia analysis and retrieval techniques provide a significant benefit over retrieval using only textual information such as from automatic speech recognition transcripts or closed captions. In interactive systems, we do find significant differences among the top systems, indicating that interfaces can make a huge difference for effective video/image search. For interactive tasks efficient interfaces require few key clicks, but display large numbers of images for visual inspection by the user. The text search finds the right context region in the video in general, but to select specific relevant images we need good interfaces to easily browse the storyboard pictures. In general, TRECVID has motivated the video retrieval community to be honest about what we don't know how to do well (sometimes through painful failures), and has focused us to work on the actual task of video retrieval, as opposed to flashy demos based on technological capabilities. 6061-10, Session 4 A color selection tool ensuring legibility of textual information on web pages S. Zuffi, Consiglio Nazionale delle Ricerche (Italy); G. B. Beretta, Hewlett-Packard Co.; C. Brambilla, Consultant (Italy) One of the issues in Web page design is the selection of appropriate combinations of background and foreground colors to display textual information. Colors have to be selected in order to guarantee legibility for different devices, viewing conditions and, more important, for all the users, including those with deficient color vision. In this work we present a tool to select background and foreground colors for the display of textual information. The tool is based on the Munsell Book of Colors; it allows the browsing of the atlas and indicates plausible colors based on a set of readability measure that can be selected from a set of criteria. In order to take into account color vision deficiencies, we focused on color selection based on luminance, and performed experiments to evaluate readability of many color combinations selected at different levels of lightness difference by means of a character counting task. As first result, our data suggest that, assigning a lightness difference of about 30 in the color selection tool could be a good strategy to address legibility issues. Our results, even if interesting, require further extensions and comparisons, that we plan to perform in the future. 6061-11, Session 4 A color interface for audio clustering visualization S. Zuffi, I. Gagliardi, Consiglio Nazionale delle Ricerche (Italy) The availability of large audio collections calls for ways to efficiently access and explore them by providing an effective overview of their contents at the interface level. We present an innovative strategy for the visualization of the contents of a database of audio records. The database is that of the Archive of Ethnography and Social History of the Lombardy Region (AESS). The website stores information concerning the oral history of the region, and is composed mainly of popular songs and other audio records describing the popular traditions handed down from generation to generation. The AESS website implements various modalities of navigation. These include the location and clustering of similar audios, that is the organization of the audio files stored in the database in groups containing files acoustically similar to each other. An innovative strategy was defined to implement a color coding of audio clusters. We exploited the visual attributes of colors to map the cluster distances in the histogram space into color distances in a perceptually uniform color space. This kind of representation can support the user in the browsing of audio clusters, allowing a rapid visual evaluation of the similarity of cluster contents without the need to listen to the audios.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
131
Conf. 6061: Internet Imaging VII
6061-12, Session 4 Interactive internet delivery of scientific visualization vis structured prerendered imagery J. Chen, San Francisco State Univ. and Lawrence Berkeley National Lab.; E. W. Bethel, Lawrence Berkeley National Lab.; I. Yoon, San Francisco State Univ. In this paper, we explore leveraging industry-standard media formats to effectively deliver interactive, 3D scientific visualization to a remote viewer. Our work is motivated by the need for remote visualization of time-varying, 3D data produced by scientific simulations or experiments while taking several practical factors into account, including: maximizing ease of use from the user's perspective, maximizing reuse of image frames, and taking advantage of existing software infrastructure wherever possible. Visualization or graphics applications first generate images at some number of view orientations for 3D scenes and temporal locations for time-varying scenes. We then encode the resulting imagery into one of two industry-standard formats: QuickTime VR Object Movies or a combination of HTML and JavaScript code implementing the client-side navigator. Since the only inputs consist of image data, a viewpoint and time stamps, our approach is generally applicable to all visualization and graphics rendering applications capable of generating image files in an ordered fashion. Our design is a form of latency-tolerant remote visualization infrastructure where processing time for visualization, rendering and content delivery is effectively decoupled from interactive exploration. Our approach trades off increased interactivity, reduced load and effective reuse of coherent frames between multiple users at the expense of unconstrained exploration. 6061-13, Session 5 Clustering and semantically filtering web images to create a large-scale image ontology S. Zinger, C. Millet, M. Benoit, G. Grefenstette, P. Hиde, P. Moлllic, Commissariat a l'Energie Atomique (France) In our effort to contribute to the closing of the "semantic gap" between images and their semantic description, we are building a large-scale ontology of images of objects. This visual catalog will contain a large number of images of objects, structured in a hierarchical catalog, allowing image processing researchers to derive signatures for wide classes of objects. We are building this ontology using images found on the web. We describe in this article our initial approach for finding coherent sets of object images. We first perform two semantic filtering steps: the first involves deciding which words correspond to objects and using these words to access databases which index text found associated with an image (e.g. Google Image search) to find a set of candidate images; the second semantic filtering step involves using face recognition technology to remove images of people from the candidate set (we have found that often requests for objects return images of people). After these two steps, we have a cleaner set of candidate images for each object. We then index and cluster the remaining images using our system VIKA (VIsual KAtaloguer) to find coherent sets of objects. 6061-14, Session 5 Ontology and image semantics in multimodal imaging: submission and retrieval Y. Bei, M. Belmamoune, F. J. Verbeek, Leiden Univ. (Netherlands) In the last decade, research activities in the life sciences are responsible for production of a tremendous amount of digital data.
Considerable part of this data is in the form of images resulting from scientific experiments and, which are often stored in different repositories without any standard approach that could facilitate scientific retrieval, analysis or exchange among them. Without accurate annotation images are not straightforwardly suitable for exchange. Knowledge about what is depicted in the image as well as specific image content is needed for understanding and thus, sharing. We introduce metadata annotation to store image content in both pixel and semantic level. Our approach focuses on structured annotation to unlock the knowledge in microscopy images. Ontology is used as a standard controlled and related vocabulary in our annotation. Different ontologies enrich annotation in different aspects. Unified ontology based annotation, will help scientists and computers understanding knowledge as present in images more precisely. It will also assure interoperability by allowing structured exchange of information among different repositories. Ontologies provide essential glue for elaborate image retrieval and assure propagation of image content by accurate linking to other resources. 6061-16, Session 6 Combining color models for skin detection F. Aldershoff, T. Gevers, H. M. Stokman, Univ. van Amsterdam (Netherlands) The choice of a colour space is of great importance for many computer vision algorithms (e.g. edge detection and object recognition). The choice of this colour space is nontrivial, since it will induce the equivalence classes in the algorithms performing the tasks. The choice is further complicated by the vast range of available colour spaces (e.g. RGB, rgb, CIE L*a*b*, HSV , etc.), each with specific qualities. Moreover the optimal colour space might be a mix of some standard colour spaces. The problem is how to automatically select the weighting to integrate the colour spaces in order to produce the best result for a particular task. In this paper we propose a method to learn these weights, while exploiting the non-perfect correlation between colour spaces of features through the principle of diversification. As a result an optimal trade-off is achieved between repeatability and distinctiveness. The resulting weighting scheme will ensure maximal feature discrimination. The method is experimentally verified for three feature detection tasks: Skin colour detection, edge detection and corner detection. In all three tasks the method achieved an optimal trade-off between (colour) invariance (repeatability) and discriminative power (distinctiveness). 6061-17, Session 6 Using context and similarity for face and location identification M. Davis, Univ. of California/Berkeley; M. A. Smith, France Telecom R&D (France); J. F. Canny, Univ. of California/Berkeley; F. W. M. Stentiford, Univ. College London (United Kingdom) This paper describes a new approach to the automatic detection of human faces and location in mobile images. The growing infrastructure of internet enabled location services and the ubiquity of position sensing cameraphones present both a challenge and opportunity for multimedia researchers. In addition to their growing global ubiquity, cameraphones offer a unique opportunity to purse new approaches to media analysis and management: namely to combine the analysis of automatically gathered contextual metadata with media content analysis to radically improve image content recognition and retrieval.
132
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6061: Internet Imaging VII
Current approaches to content-based image analysis are not sufficient to enable retrieval of cameraphone photos by high-level semantic concepts, such as who is in the photo or what the photo is actually depicting. In this paper, new methods for determining image similarity are combined with analysis of automatically acquired contextual metadata to produce location information. For faces, context-aware, collaborative filtering, machine-learning techniques that leverage automatically sensed and inferred contextual metadata together with computer vision analysis of image content to make accurate predictions about the human subjects depicted in cameraphone photos. Our database consists of over 1200 images that were collected on cameraphones and annotated to provide context features such as outdoor/indoor setting and time of capture. Most of the images were taken in natural settings with limited frontal pose. For location, a model of Cognitive Visual Attention (CVA) matches large numbers of pairs of pixel groups (forks) taken from two patterns under comparison. We achieve a significant reduction from a 70% error rate from color histogram and CVA image analysis methods for determining the location of photo content, to a 45% error rate using contextual metadata alone, to a 33% error rate achieved by combining contextual metadata with CVA image analysis. For faces, we apply Sparse-Factor Analysis (SFA) to both the contextual metadata gathered in th database and the results of PCA (Principal Components Analysis) of the photo content to achieve a 60% face recognition accuracy of people depicted in our cameraphone photos, which is 40% better than media analysis alone. The margins in precision/recall among the different methods are quite large. Context+Vision does better than any individual predictor. Its initial precision is about 60% and is fairly flat across the recall range, as seen in the figure below. This paper has shown that the combination of attributes derived from both contextual metadata and image processing produces a measure that can indicate the faces in a photo and the location at which photos were taken, much better than image analysis or context alone. 6061-18, Session 6 Skin segmentation using multiple thresholding F. Gasparini, R. Schettini, Univ. degli Studi di Milano-Bicocca (Italy) The segmentation of skin regions in color images is a preliminary step in several applications. Many different methods are available in the literature. The simplest methods build what is called an "explicit skin cluster" classifier which expressly defines the boundaries of the skin cluster in certain color spaces. The main difficulty in achieving high skin recognition rates, and producing the smallest possible number of false positive pixels, is that of defining accurate cluster boundaries through simple, often heuristically chosen, decision rules. We apply a genetic algorithm to determine the boundaries of the skin clusters in multiple color spaces. To quantify the performance of these skin detection methods, we use recall and precision scores. A good classifier should provide both high recall and high precision, but generally, as recall increases, precision decreases. Consequently, we adopt a weighted mean of precision and recall as the fitness function of the genetic algorithm. The weighting coefficients can be chosen to favor either high recall or high precision, or to satisfy a reasonable tradeoff between the two, depending on application demands. To train the genetic algorithm and test the performance of the classifiers applying the suggested boundaries, we use the large and heterogeneous Compaq database.
6061-19, Session 7 Integration of multimedia contents and elearning resources in a digital library M. Pascual, N. Ferran, J. Minguillуn Alfonso, Univ. Oberta de Catalunya (Spain) Distance education in virtual e-learning environments permits an intensive use of new technologies, especially in the field of design, creation and management of multimedia contents. The use of multimedia resources, either as learning tools in virtual environments or as basic pieces of multimedia repositories, allows us an improvement in the process of teaching and learning contents of audiovisual nature. In this sense, teaching contents with a clear multimedia structure, such as several subjects in Studies of Audiovisual Communication, for example, but also Multimedia or Humanities Studies, requires learning tools with two desired characteristics: first, each course follows an activity oriented structure using a temporal framework, partially fixed but flexible, and second, content personalization capabilities are needed to create adaptive courses depending on user preferences and background, but also to minimize course obsolescence by means of semiautomated content update. Therefore, content reusability within a course is an important issue related to course quality and management, one of the main goals of this ongoing project. This paper describes the integration between documents according to standards for multimedia content distribution such as MPEG-7, and other learning resources designed using e-learning standards such as LOM, for example. This is performed within the context of a virtual e-learning environment and a digital library, which uses the Dublin Core standard for metadata or MARC format among other standards. The need for metadata integration across multimedia, elearning and library standards becomes a key factor for ensuring a proper content management and retrieval by teachers, researchers and students, the users of the digital library. These standards are not orthogonal, so an overlap minimization and the appropriate extensions must be performed to ensure a proper content tagging, mainly to fulfill the main project goal, basically that all the resources in the digital library can be browsed and searched as a large repository of multimedia contents which uses structured metadata for the syntactic and semantic description of all the resources, which can be used in different scenarios. Metadata provide controlled and structured descriptions of learning resources through searchable access points such as title, author, date, location, description and subject, but can also provide interpretative information on the potential education application of resources or include described information about the relationships with other resources. The learning resources form a hierarchical structure when are combined to create courses using learning objects as the basic pieces. These courses have a flexible structure that allows teachers to adapt their content on-the-fly, depending on any particular teaching requirement. On the one hand, we define the appropriate mappings between multimedia, e-learning and library standards in a two-stage approach: first, a set of common metadata is identified in all three standards for supporting basic browsing and searching capabilities and, second, the appropriate extensions usingontologies are created to minimize metadata overlap and inconsistencies. The use of XML based languages such as Resource Description Framework (RDF) and Web Ontology Language (OWL) is also addressed. On the other hand, the specific information needed to describe e-learning usage of such multimedia contents, related to pedagogical aspects, needs also to be specified. Finally, all the elements in the digital library are under an intellectual property rights management policy, which determines the possible contexts and scenarios of use. The digital library tracks all searching and browsing actions with two main goals: first, to improve such
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
133
Conf. 6061: Internet Imaging VII
capabilities by means of collaborative filtering and annotation, and second, to ensure that all contents that are part of a course and the users of such course (both teachers and students) have the appropriate rights to access them. 6061-20, Session 7 Selecting the kernel type for a web-based adaptive image retrieval systems (AIRS) A. Doloc-Mihu, V. V. Raghavan, Univ. of Louisiana at Lafayette The goal of this paper is to investigate the selection of the kernel for a Web-based AIRS. Using the Kernel Perceptron learning method, several kernels having polynomial and Gaussian Radial Basis Function (RBF) like forms (6 polynomials and 6 RBFs) are applied to general images represented by color histograms in RGB and HSV color spaces. Experimental results on these collections show that performance varies significantly between different kernel types and that choosing an appropriate kernel is important. 6061-22, Session 7 Medical validation and CBIR of spine x-ray images over the Internet S. K. Antani, J. Cheng, J. L. Long, L. R. Long, G. R. Thoma, National Library of Medicine As found in the literature, most Internet-based prototype ContentBased Image Retrieval (CBIR) systems focus on stock photo collections and do not address challenges of large specialized image collections and topics such as medical information retrieval by image content. Even fewer have medically validated data to evaluate retrieval quality in terms of precision and relevance. To date, our research has reported over 75% relevant spine X-ray image retrieval tested on 888 validated vertebral shapes from 207 images using our prototype CBIR system operating within our local network. As a next step, we have designed and developed an Internet-based medical validation tool and a CBIR retrieval tool in MATLAB and JAVA that can remotely connect to our database. The retrieval tool supports hybrid text and image queries and also provides partial shape annotation for pathology-specific querying. These tools are initially developed for domain experts, such as radiologists and educators, to identify design issues for improved workflow. This article describes the tools and design considerations in their development. 6061-23, Session 7 The integration of cartographic information into a content management system M. M. Furnari, C. I. D. Noviello, Consiglio Nazionale delle Ricerche (Italy) A corporate information system needs to be as accessible as library content, which implies to organize the content in a logical structure, categorizing it, and using the categories to add metadata to the information. Content Management System (CMS) are an emerging kind software component that manages content, usually making a large use of the web technologies, whose main goals are to allow easy creation, publishing and retrieval of content to fit business needs. The focus of this paper is to describe how we integrated "map" metaphor into a CMS. Where maps are symbolic information and rely on the use of a graphic sign language. A characteristic feature of maps is that their design has traditionally been constrained by the need to create one model of reality for a variety of purposes. The map's primary role as a communication medium involves the application of processes such as selection, classification, displacement, symbolization and graphic exaggeration. A model of
the infrastructure is presented and the current prototype of the model is briefly discussed together the currently deployed environment for the cultural heritage information dissemination. 6061-36, Session 7 FaceLab: a tool for performance evaluation of face recognition strategies L. Caflisch, Comerson s.r.l. (Italy); A. Colombo, C. Cusano, R. Schettini, F. Tisato, Univ. degli Studi di Milano-Bicocca (Italy) This paper presents FaceLab, an innovative, open environment created to evaluate the performance of face recognition strategies. It simplifies, through an easy-to-use graphical interface, the basic steps involved in testing procedures such as data organization and preprocessing, definition and management of training and test sets, definition and execution of recognition strategies and automatic computation of performance measures. The user can extend the environment to include new algorithms, allowing the definition of innovative recognition strategies. The performance of these strategies can be automatically evaluated and compared by the tool, which computes several performance measures for both identity verification and identification scenarios. 6061-24, Session 8 Enhanced video display and navigation for networked streaming video and networked video playlists S. G. Deshpande, Sharp Labs. of America, Inc. In this paper we present an automatic enhanced video display and navigation capability for networked streaming video and networked video playlists. Our proposed method uses Synchronized Multimedia Integration Language (SMIL) as presentation language and Real Time Streaming Protocol (RTSP) as network remote control protocol to automatically generate a "enhanced video strip" display for easy navigation. We propose and describe two approaches - a smart client approach and a smart server approach. We also describe a prototype system implementation of our proposed approach. 6061-25, Session 8 3D display technique for moving pictures from web cameras using screen pixel accessing T. Hasegawa, T. Namiki, H. Unno, K. Uehira, H. Kasuga, K. Yanaka, Kanagawa Institute of Technology (Japan) This paper presents a technique to display real-time 3-D images captured by web cameras on the stereoscopic display of a personal computer (PC) using screen pixel access. Images captured by two side-by-side web cameras are sent through the Internet to a PC and displayed in two conventional viewers for moving images. These processes are carried out independently for the two cameras. The image data displayed in the viewer are in the video memory of the PC. Our method uses this video-memory data, i.e., the two webcamera images are read from the video memory, they are composed as a 3-D image, and then it is written in the video memory again. A 3-D image can be seen if the PC being used has a 3-D display. We developed an experimental system to evaluate the feasibility of this technique. The web cameras captured up to 640 x 480 pixels of an image, compressed it with motion JPEG, and then sent it over a LAN. Using an experimental system, we evaluated that the 3-D image had almost the same quality as a conventional TV image by using a broadband network like ADSL.
134
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6061: Internet Imaging VII 6061-28, Session 9 Dynamic conversion between XML-based languages for vector graphics A. Di Iorio, F. Vitali, G. Zonta, Univ. degli Studi di Bologna (Italy) Vector graphics is increasingly gaining importance within the Word Wide Web community, because it allows users to create images that are easily manageable, modifiable and understandable. Two formats play a leading role among the languages for vector graphics: SVG and VML. Achieving a complete interoperability between these two languages means providing users a complete support for vector images across implementations, operating systems and media. Even automatic conversions between raster and vector graphics provide users a lot of benefits and are worth further investigation and support. In this paper we describe VectorConverter, a tool developed at the University of Bologna that allows easy, automatic and reasonably good conversion between two vector graphic formats, SVG and VML, and one raster format, GIF. This tool makes good translations between languages with very different functionalities and expressivity, by applying translation rules, approximation and heuristics. VectorConverter is composed of a few XSLT files to manage the conversion between vector formats, and some PHP scripts that work on raster images. A high-level discussion about implementation details, open issues and future developments of VectorConverter is provided as well. 6061-29, Session 9 Bezier curves approximation of triangularized surfaces using SVG G. Messina, STMicroelectronics (Italy); E. Ingra, S. Battiato, G. Di Blasi, Univ. di Catania (Italy) This paper presents a technique to convert surfaces, obtained using a Data Dependent Triangulation (DDT), in Bйzier Curves through a Scalable Vector Graphics (SVG) file format. The DDT replaces the input image with a set of triangles according to a specific cost function able to detect the edge details. On the other hand the DDT produces a number of triangles larger than the number of pixels. Although the quality achieved in this way is rather good the size of the resulting files may be very large. To reduce the amount of data the algorithm extracts from the DDT only the boundaries triangles; then the triangles are synthesized as single points using their estimated barycenters. These barycenters are connected together by following the boundaries along the wind rose directions. After the areas are created, the conversion to Bйzier Curves is performed by using the resulting path as control points. A simplification of the curves is then applied by removing useless points. Finally the surfaces are sorted using an approximation of the area and saved in SVG format. The proposed technique is compared with other raster to vector conversion methods and software showing good performances in terms of perceptual and measured quality.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
135
Conf. 6062: Spectral Imaging: Eighth International Symposium on Multispectral Color Science Monday-Tuesday 16-17 January 2006 Part of Proceedings of SPIE Vol. 6062 Spectral Imaging: Eighth International Symposium on Multispectral Color Science
6062-01, Session 1 Hyperspectral imaging of sulfate evaporite deposits in Western Australia and on Mars A. J. Brown, Macquarie Univ. (Australia) and Australian Ctr. for Astrobiology (Australia); T. J. Cudahy, Commonwealth Scientific & Industrial Research Organisation (Australia) The European hyperspectral imaging instrument Observatoire pour la Minйralogie, l'Eau, la Glace et l'Activitй (OMEGA) has been in operation around Mars since early 2004. OMEGA has constructed imaging maps covering almost the entire Martian surface. OMEGA has returned evidence of surficial sulfate deposits at several locations around the Martian globe.The presence of sulfates on the Martian surface has important links with past water and possible life on Mars. On Earth, sulfates most commonly form in dry lake evaporite basins, many examples of this type of deposit are in evidence in arid Western Australia. A number of these dry lakes criss-cross the ancient (late Archean) Yilgarn Craton. The ultramafic-mafic volcanic flows in this region make the Yilgarn a good analog for the volcanic flood basalts of Mars. In 2004, a hyperspectral imaging survey of the Yilgarn Craton was carried out using the airborne HyMap instrument. We have analysed this hyperspectral coverage of the evaporite deposits of the Yilgarn GGT and found large deposits of gypsum in evidence. Using an analysis method based on curve fitting of individual spectra in the dataset, we have compared the results for the Martian North Polar region with the arid Western Australian evaporite deposits. 6062-02, Session 1 Multispectral imaging determination of pigment concentration profiles in meat C. Sбenz, B. Hernбndez, C. Alberdi, S. Alfonso, M. Berrogui, J. M. Diсeiro, Univ. Publica de Navarra (Spain) The possibility of using multispectral techniques to determine the concentration profiles of myoglobin derivatives as a function of the distance to the meat surface during meat oxygenation is demonstrated. Reduced myoglobin (Mb) oxygenated oxymyoglobin (MbO2) and oxidized Metmyoglobin (MMb) concentration profiles are determined with a spatial resolution better than of 0.01235 mm/ pixel. Pigment concentrations are calculated using (K/S) ratios at isobestic points (474, 525, 572 and 610 nm) of the three forms of myoglobin pigments. This technique greatly improves previous methods, based on visual determination of pigment layers by their color, which allowed only estimations of pigment layer position and width. The multispectral technique avoids observer and illumination related bias in the pigment layer determination. 6062-03, Session 1 Visualization of the human face skin moisturizing-ability by spectroscopic imaging using two near-infrared bands H. Iwasaki, K. Miyazawa, S. Nakauchi, Toyohashi Univ. of Technology (Japan) The skin's ability to retain moisture, which is hereafter referred as skin moisturizing-ability, is one of the important factors in skin health. Skin defends the biological tissue from the outside influences, skin sebum and moisture especially play an important role in that protection. Near-infrared (NIR) spectroscopic imaging has
recently been capable of detecting changes in skin hydration of the forearms. However, face skin hydration has not been measured, and the moisture-related sebum has not been paid attention to, even though the face is important from the cosmetic point of view. This study, therefore, aims to measure and visualize the spatial distribution of skin moisturizing-ability of the face by NIR spectroscopic imaging. The NIR spectral imaging system consists of two interference filters (1060 and 1450nm) mounted on a filter wheel and a NIR camera with indium-gallium arsenide array sensor. We measured face skins with/without moisturizing lotion and the areas where moisturizing lotion was applied were successfully displayed by subtracting two absorbance images measured at different wavelength bands. It was also found that the glabella and nose have strong moisturizing-ability because of sebaceous glands. This technique can be applied to the functional assessment of face skin moisturizer in medicine and cosmetics. 6062-04, Session 2 Image processing techniques for detection of buried objects in infrared images A. Ceron-Correa, Univ. Militar Nueva Granada (Colombia); O. L. Lopera, Royal Military Academy (Belgium) and Univ. de Los Andes (Colombia) This paper describes the features of infrared thermography and its application to humanitarian demining in the world and also the influencing factors on its application in a country like Colombia which suffers hardly the antipersonnel mines problem. Infrared image processing methods and results of tests done in different sites in the country are showed. The IR cameras are sensitive passive devices; they are capable to detect infrared radiation of buried objects in specific situations. These sensors are based in the alteration of heat flux, occasioned by the presence of strange bodies buried in the soil. This situation produces differences in the infrared radiation emission from objects in the location because they hold a different amount of heat than in its surroundings. The images are affected for different factors in special when the soil is composed of stones which can give a great number of false alarms; a method for estimating this feature named granulometry is used in this project. Finally, a method for the detection of the signature of a buried object is used with successful results. This method uses a combination of morphology mathematic techniques. 6062-05, Session 2 Spectral estimation of made-up skin color under various conditions M. Doi, R. Ohtsuki, S. Tominaga, Osaka Electro-Communication Univ. (Japan) A method is proposed for estimating the spectral reflectance of made-up skin color under various conditions including the undesirable colored skin. The color of dark spot is caused by increasing the component of melanin. The reddish skin is caused by the increase of hemoglobin. Our method uses the Kubelka-Munk theory to calculate the surface spectral reflectance human skin. This theory calculates the reflectance and transmittance of the light passing through a turbid medium from the absorption and scattering of the medium. The spectral reflectance of made-up skin is estimated by adjusting parameters of the thickness of the makeup layer. The proposed estimation method is evaluated on an
136
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6062: Spectral Imaging: Eighth International Symposium on Multispectral Color Science
experiment in detail. First, we measure the spectral reflectance of facial skin under the three conditions of normal skin, undesirable skin, and made-up skin. The undesirable skin includes stain, suntan or ruddy skin. The made-up skin means the skin with foundation on the normal skin, the stain, the suntan and the ruddy skin. Second, we estimate the spectral reflectance of made-up skins from the reflectance of bare skins and optical characteristics of foundations. Good coincidence between the estimated reflectance and the direct measurement shows the feasibility of the proposed method. 6062-06, Session 2 MODIS versus ASTER water classification C. Alecu, S. Oancea, National Meteorological Administration (Romania); E. Bryant, Dartmouth College Moderate Resolution Imaging Spectroradiometer (MODIS) and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) are multi-spectral sensors embarked on the EOS AM-1 (TERRA) satellite platform. Both sensors operate in different spectral bands, but also with different pixel resolutions. Due to the constraint that high spatial resolution satellite images are low temporal resolution, there exists a need for a reliable method to obtain accurate information from medium resolution data. The overall goal of this paper is to classify MODIS data to get an estimation of water surface area, very useful in the flood management for the decision makers at all levels. To develop the classification technique, the strategy was to obtain MODIS and ASTER data acquired at the same time over the same location, and use the ASTER data as "ground truth". Two lakes in the Bihor County of Romania were chosen as test area. Water masks were created from ASTER and MODIS data and then superimposed, taking into account that a MODIS pixel contains 400 ASTER pixels. Each MODIS pixel was converted as ratio of water/non-water, based on the ASTER classification. The water surface area, as measured from the MODIS classification, was about 16% more than the ASTER ground truth-value. 6062-07, Session 2 Improving multivariate curve resolution analysis performance when applied to fluorescence hyperspectral biological imaging H. D. T. Jones, E. Thomas, D. M. Haaland, J. A. Timlin, M. B. Sinclair, Sandia National Labs. Hyperspectral imaging is becoming a powerful new tool to image biological cell and tissue samples. In a hyperspectral image, hundreds of wavelengths representing whole spectral regions are obtained at each pixel (or voxel for 3-dimensional images). We are using multivariate Curve Resolution (MCR) to analyze the collected image data. MCR is a powerful analysis technique for hyperspectral images since it has the ability to extract pure-component spectra from the spectral images and provide relative quantitative determinations of each component for each pixel in the image. The MCR algorithm performs best when the data are unique, (i.e., orthogonal) in either the spectral or spatial domain. However, this condition is often not met in experimental datasets. In this presentation, I will discuss the use of data preprocessing techniques to select the image pixel spectra to include in the MCR analysis in order to improve the concentration orthogonality of the pixels used in the analysis. The goal is to have the MCR result converge to the most appropriate solution when random numbers are used to initiate the MCR process. The methods to select the most appropriate pixels for MCR analysis will be demonstrated using hyperspectral images of photosynthetic cyanobacteria.
6062-08, Session 3 Estimating reflectance parameters from saturated spectral images S. Li, Y. Manabe, K. Chihara, Nara Institute of Science and Technology (Japan) Since commercial image detectors, such as charge-coupled device (CCD) cameras, have a limited dynamic range, it is difficult to obtain images that really are unsaturated, as a result of which the reflectance parameters may be inaccurately estimated. To solve this problem, we describe a method to estimate reflectance parameters from saturated spectral images. We separate reflection data into diffuse and specular components at 5-nm intervals between 380nm and 780nm for each pixel of the spectral images, which are captured at different incident angles. To estimate the specular reflectance parameters from the specular components, we transform the Torrance-Sparrow equation to a linear form. We estimate specular parameters for intensity of the specular reflection and standard deviation of the Gaussian distribution, using the least squares method from unsaturated values of the specular components, assuming Fresnel reflectance is constant. And then, to refine estimation of those specular parameters, we estimate the refractive index of the surface for dielectric materials in terms of the Fresnel equation. We carried out experiments with measured data, and with simulated specular components at different saturation levels, generated according to the Torrance-Sparrow model. Our experimental results reveal that the diffuse and specular reflectance parameters are estimated with high quality. 6062-09, Session 3 Influence of the recovery method in the optimum sensors for spectral imaging of skylight M. A. Lopez-Alvarez, J. Hernandez-Andres, J. L. Nieves, J. Romero, Univ. de Granada (Spain) We show the similarities and differences between the optimum sensors found for recovering skylight spectra from noisy broadband sensor data with four different spectral estimation methods. We also study the accuracy obtained in the spectral recoveries of skylight with each method. The Maloney-Wandell and Wiener methods, although very different mathematically, present a similar behaviour in the spectral profile of the optimum sensors and the quality of the recoveries. The Imai-Berns method improves a little upon these two methods when the noise present in the system is high. The optimum sensors obtained with the Shi-Healey method are very peculiar, they seem to be equally spaced in the visible and are very narrow-band, which indicates that they could easily be constructed using a Lyquid Crystal Tunable Filter. For the Maloney-Wandell and the Imai-Berns methods the best choice is to use 3 basis vectors when recovering skylight spectra from three-sensor responses. The Shi-Healey method has proved to be the best for the task of recovering skylight spectra from the responses of three sensors, although it is extremely slow when using a large training set. 6062-10, Session 4 Demosaicking methods for multispectral cameras using mosaic focal plane array technology G. A. Baone, H. Qi, The Univ. of Tennessee This paper focuses on the introduction of mosaic focal plane array technology to multispectral image acquisition systems. The motivation to use mosaic focal plane arrays for multispectral image acquisition comes from the commercial digital color cameras. The digital color cameras use an array of sensors to sense incoming light
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
137
Conf. 6062: Spectral Imaging: Eighth International Symposium on Multispectral Color Science
such that only one spectral band is registered per pixel location. This helps in reducing the cost and size of the equipment and at the same time increasing the robustness and improving the registration capability of the equipment. The final (full color) image is formed by estimating the missing spectral bands at each pixel location. This process is called demosaicking. In this paper, we deal with the demosaicking aspect of the multispectral image acquisition systems that use mosaic focal plane arrays. Existing demosaicking techniques have been appropriately extended to the multispectral case. To incorporate external noise and degradations during the acquisition process, we use a Maximum a-Posteriori probability (MAP) based approach that treats the demosaicking problem as a classic case of image restoration. This MAP based approach successfully performs demosaicking by simultaneously removing any existing noise and degradations. Experiments have been performed on a set of seven-band multispectral images and various metrics have been developed to compare different demosaicking techniques. 6062-11, Session 4 Estimation of noise variance of a multispectral image acquisition system N. Shimano, Kinki Univ. (Japan) The noise present in a color image acquisition system influences the accuracy of the estimated colorimetric values and the accuracy of the recovered spectral reflectances of objects being imaged through the use of sensor responses. Estimation of the noise levels in the devices is important for the accurate acquisition of colorimetric or spectral information. This work addresses the problem for the determination of noise variances in multispectral image acquisition systems. In the present paper four different models, i.e., the first two models are based on the recovered spectral reflectances by the Wiener filter and the second two models are based on the statistical analysis of sensor responses, for the determination of the noise variances are briefly reviewed, and they were applied to a multispectral image acquisition system in detail. From the experimental results, it is confirmed that the estimated noise variances by the model which uses recovered spectral reflectances are more accurate than those estimated noise variances by the statistical analysis of sensor responses. The proposal is very important for the recovery of the surface reflectance spectra of objects being imaged and also useful for the evaluation of a set of sensors. 6062-12, Session 4 Multispectral stand-off imaging with midinfrared semiconductor lasers Y. Wang, Y. Wang, H. Q. Le, Univ. of Houston Multi-spectral laser imaging can be used for target discrimination, classification, and identification based on object spectral signatures. This paper describes a development of semiconductor-laser-based multi-spectral and polarimetric imaging, using near-IR and mid-IR lasers. Key issues and aspects of this technology are discussed, including the fundamental issues of multi-spectral imaging and scattering phenomena, and the system engineering approach for multi-wavelength scaling. The near-IR study employed 7 wavelengths from 0.635-1.55 µm, and demonstrated fundamental aspects of wavelength- and polarization-dependence effects in imaging and scattering. Stokes vector imaging was shown to reveal significant information of the targets. The mid-IR study employed 4 wavelengths from 3.3-9.6 µm, and was applied to diverse targets consisting of natural and man-made materials and household objects. It was shown capable to resolve and distinguish small spectral differences among targets, e. g. colorless objects in the
visible were shown with "colorful" signatures in the mid-IR. For system engineering design, a key feature is the system architecture that employs wavelength-division-multiplexing for accurate and high spectral fidelity, and the scalable CDMA network approach with multiple transmitters and receivers for efficient measurements. The results suggest that multi-spectral laser imaging can be a unique and powerful technology for wide ranging applications. 6062-13, Session 4 Designing flat-bed scanning system for spectral and glossiness recording T. Takiguchi, S. Abe, T. Makino, N. Tsumura, T. Nakaguchi, Chiba Univ. (Japan); F. Nakaya, H. Ichikawa, Y. Minato, Fuji Xerox Co., Ltd. (Japan); K. Miyata, National Museum of Japanese History (Japan); Y. Miyake, Chiba Univ. (Japan) In this paper, we propose a flat-bed scanner system to record spectral and glossiness information for various sheets like objects. In the proposed system, five filters are used to acquire the multiband images of the object. The spectral reflectance image can be estimated and be recorded from the multi-band images. The glossiness of the object is recorded as two images taken by the different geometries about illuminant which is from 45 degrees, and from 0+ degrees respectively when the averaged normal vector of the sheet like object is defined as 0 degree. We performed two types of computer simulation by using the two images to reproduce various appearance of recorded object. As the first simulation, the various appearances of image are reproduced through a weighted linear combination of the two images. As the second simulation, the normal vector distribution of the object is estimated from the image taken by 45 degree illuminant. By using this normal vector distribution, roughness of the object is estimated from the image taken by 0 + degree illuminant. The normal vector and estimated roughness are used to reproduce the various appearance of the object under arbitrary illuminant. 6062-15, Session 4 Color measurements with colorimetric and multispectral imaging systems M. de Lasarte, M. Vilaseca, J. Pujol, M. Arjona, Univ. Politиcnica de Catalunya (Spain) This work is focused on the study and comparison of the performance for color measurements of different systems based on optoelectronic imaging sensors. In this context we use both colorimetric, that is, with only three acquisition channels, and multispectral configurations, with more spectral bands, in order to measure the color associated to each pixel of the captured scene. The used configurations consist of a 12-bit depth monochrome CCD camera (QImaging) with a high spatial resolution attached to an objective lens (Nikon), and several broadband filters placed between them. For the colorimetric configurations with a three channel set-up an RGB tunable filter is used, and in the multispectral-based systems a motorized wheel with seven interference filters is placed between the camera and the lens instead. The whole systems have been spatially characterized using a flat-field correction. From the raw digital responses measured with the different configurations we calculated the X, Y and Z tristimulus values using four different methodologies. The first method (colorimetric), which has been studied in previous work, provides the XYZ tristimulus values by means of an absolute spectral and colorimetric characterization of the system, which permits us to obtain the spectral sensitivities associated to the sensor and the colorimetric profile between the RGB space of the device and the CIE-1931 XYZ standard space. The second (colorimetric) permits us to directly transform the RGB digital signals to XYZ values performing a simple mathematical fitting between both sets, such as a minimum least squares regression. The third method (multispectral) uses the same principle
138
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6062: Spectral Imaging: Eighth International Symposium on Multispectral Color Science
as the second one but taking into account the seven digital signals associated to the multispectral bands instead of using only the RGB values. The last proposed methodology (multispectral) initially performs the reconstructions of the spectra associated to each pixel using different methods such as the Moore-Penrose pseudo-inverse and finally computes the XYZ tristimulus values from the spectral information provided. The different proposed systems were experimentally tested using the Gretagmacbeth ColorChecker chart and a tele-spectracolorimeter PhotoResearch PR-650 was also used in order to measure the associated true colors of the patches. The results obtained so far show the improvement on the color measurements with the multispectral-based systems, which are more accurate than the ones obtained with the systems with only three acquisition channels. The study also shows the worse performance of the first colorimetric methodology due to the great amount of errors carried along the calculations involved in the spectral and colorimetric characterizations. Finally, the fourth methodology is preferred if spectral information is useful in addition to the colorimetric parameters. The developed systems, whichever configuration is used, may be integrated as an intelligent sensor in automatic manufacturing cells and allows either color measurements using a single register with customizable spatial resolution and offering a fairly higher spatial resolution when compared with standard systems for color measurements. 6062-16, Session 5 High-fidelity video and still-image communication based on spectral information: natural vision system and its applications M. Yamaguchi, Tokyo Institute of Technology (Japan); H. Haneishi, Chiba Univ. (Japan); H. Fukuda, J. Kishimoto, H. Kanazawa, M. Tsuchida, R. Iwama, National Institute of Information and Communications Technology (Japan); N. Ohyama, Tokyo Institute of Technology (Japan) [Background and Purpose] The reproduction of natural color is one of the key issues in visual communication systems for electronic commerce, telemedicine, and digital museum, in addition to the technologies of high resolution or large screen displays. To breakthrough the limitation of current color imaging systems, which are based on RGB, "Natural Vision (NV)" project has been exploiting the technologies for spectrum-based color reproduction, multispectral imaging, and multiprimary displays, until the project is completed on March 2006 as scheduled. Researchers from two academic institutes (Tokyo Institute of Technology and Chiba University) and industry (NTT Data, Olympus, NTT, etc.) are participated in the NV project, and have addressed the spectrum-based color reproduction technologies for still image (1999-2003) and motion picture (2001-2006). This paper summarizes the results of the seven years activity of NV project, including the multispectral and multiprimary imaging technologies and the experimental investigations on the applications to medicine, digital archives, electronic commerce, and computer graphics. [Technologies] In the NV system [1], spectral information is acquired by multispectral image capture to accurately reproduce the colors under arbitrary illuminant, which may be different from the imagecapturing environment, as if the object is placed at the front of the observer. In the image display, the reproducible range of color (color gamut) is expanded by multiprimary color displays, in which more than three primary colors are mixed to represent color images. NV project has developed 16-band multispectral camera for still image capturing, 6-band HDTV camera for motion picture acquisition [2], 6-primary color projection displays and 4-primary color flat panel LCD for wide-gamut color reproduction. For the
interchange between the image input and display systems with arbitrary number of channels, i.e., multispectral and multiprimary systems, the spectrum-based color reproduction system [1] and the natural vision data file format are proposed. The natural color reproduction in the real-time video system is also realized with the use of color conversion device, in which the newly developed multiprimary color conversion algorithm is implemented. We have shown that the multispectral and multiprimary imaging system also enables to reduce the effect of observer metamerism, while the color difference due to the observer metamerism cannot be ignored in conventional trichromatic systems under colorimetric color reproduction. [Experimental results and conclusion] Experiments to investigate the feasibility and the applicability of NV system were carried out in the medical application such as dermatology and pathology, the color image capturing for graphic art printing, digital archives of artworks, cultural heritages [3] and natural scenes, textile and fashion industry applications with multispectral BRDF measurement, and the art expression using wide-gamut computer graphic systems. The effectiveness of the developed system has been confirmed through the experiments with the potential users of spectrum-based color reproduction. [References] [1] M. Yamaguchi, T. Teraji, K. Ohsawa, T. Uchiyama, H. Motomura, Y. Murakami, and N. Ohyama, "Color image reproduction based on the multispectral and multiprimary imaging: Experimental evaluation," Proc. SPIE, vol.4663 (2002) 15-26 [2] K. Ohsawa, T. Ajito, H. Fukuda, Y. Komiya, H. Haneishi, M. Yamaguchi, and N. Ohyama, "Six-band HDTV camera system for spectrum-based color reproduction," J. Imaging Science and Technology, vol.48, no.2, (2004) 85-92 [3] H. Fukuda, M. Yamaguchi, H. Haneishi, and N. Ohyama, "Development of 16-bands multispectral image archiving system," Proc. SPIE, Vol.5667 (2005) 136-145 6062-17, Session 5 Encoding of spectra for multiple observers and multiple illuminants T. Boosmann, RWTH Aachen (Germany) In this paper, the encoding of spectra is studied considering a variety of different observers on the one hand and a large set of standardized and non-standardized illuminants on the other. A number of 24 different observers has been defined including the CIE 1931 standard observer and the CIE 1964 supplementary standard observer as well as standard deviators. Others are selected with respect to largest differences from measurements published by Stiles and Burch. In addition, different illuminants are applied. Altogether, 52 illuminants are considered for reproducing color stimuli including such as D50, D55, D65, D75, A, B, C, E, and F1 to F12 as well as measured ones. So, this set contains illuminants with uniform spectral radiating power on the one hand and spiky power distribution on the other. A number of different encoding methods has been tested by calculating color encoding errors for all combinations of observers, illuminants and a representative set of test spectra assembled from data of Vrhel and Pointer. The encoding methods considered are based on expansions into basis functions derived from original or pre-distorted test spectra, Fourier and sine series as well as expansions into modifications of the latter ones. Maximum as well as average errors are presented.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
139
Conf. 6062: Spectral Imaging: Eighth International Symposium on Multispectral Color Science
6062-18, Session 5 Spectral-based color reproduction for print illuminated by image projector K. Ueda, S. Yamamoto, N. Tsumura, T. Nakaguchi, Y. Miyake, Chiba Univ. (Japan) We propose a new color matching method between a proof and a target print by using a projection display system based on the spectral color reproduction. In this method, a color of proof is corrected by synthesizing a projection image which is calculated to minimize the color difference between each print. The radiance of the proof and the target print are calculated by using the reflectance of the prints and the radiance from the projector, and we use the method based on the XYZ tristimulus values (colorimetric method) and the spectral values (spectral-based method). We compared the color difference between the colorimetric method and spectralbased method. The average color difference was 4.00 by using the colorimetric method. On the other hand, the average color difference was 2.13 by using the spectral-based method. From these results, we concluded that the spectral-based method is more effective than the colorimetric method to perform the accurate color reproduction in synthesizing the projection color and the proof. We believe that the proposed method was significant to simulate and control the color of proof in the printing industry. 6062-19, Session 5 Spectral-based optimization of screen images for industrial product presentation L. Hдrkцnen, J. B. Martinkauppi, H. T. Laamanen, M. Hauta-Kasari, Joensuu Yliopisto (Finland); P. Huhtelin, P. Horttanainen, Tulikivi Oyj (Finland) This paper presents a process used to optimize images for an industrial show room and its results. The selected show room is a quite typical one and therefore its light conditions are not very controllable. The projector used in the room is not a high quality one but its quality is quite standard for this kind of use. Color gamut (i.e. colors that the projector can produce) is good when compared against NTSC and sRGB gamut. The optimization of images is done using metameric reproduction and to do this we measure spectral information of the product (the soap stone tiles provided by the company, Tulikivi), projector and the illumination at the show room. The color temperature of the projector and prevailing room illumination were very different. Because they couldn't be adjusted to the same value this caused additional challenges. However, the spectral characteristic of the red channel of the projector was surprising: the range of possible red values was narrower than the green and blue range. This caused some limitations which needed to be taken into account in calculating the optimal images: optimal images can have either full contrast range with a reddish tint or correct hue with narrower contrast range. 6062-20, Session 5 Spectral gamuts and spectral gamut mapping M. R. Rosen, Rochester Institute of Technology; M. W. Derhak, Onyx Graphics All imaging devices have two gamuts: the stimulus gamut and the response gamut. The response gamut of a print engine is typically described in CIE colorimetry units, a system derived to quantify human color response. More fundamental than colorimetric gamuts are spectral gamuts, based on radiance, reflectance or transmittance units. Spectral gamuts depend on the physics of light or on how materials interact with light and do not involve the
human's photoreceptor integration or brain processing. Methods for visualizing a spectral gamut raise challenges as do considerations of how to utilize such a data-set for producing superior color reproductions. Recent work has described a transformation of spectra reduced to 6-dimensions called LabPQR. LabPQR was designed as a hybrid space with three explicit colorimetric axes and three additional spectral reconstruction axes. In this paper spectral gamuts are discussed making use of LabPQR. Also, spectral gamut mapping is considered in light of the colorimetric-spectral duality of the LabPQR space. 6062-21, Session 6 A technique for detecting metameric color areas for investigation of historical materials K. Miyata, National Museum of Japanese History (Japan); H. T. Laamanen, T. Jaaskelainen, M. Hauta-Kasari, J. P. Parkkinen, Joensuu Yliopisto (Finland) Historical materials tend to be very fragile, and thus, they need to be preserved from further degradations in future. Spectral reflectance includes a variety of objective and device independent information, which is sufficient for analysis of the materials for historical research and advanced investigation purposes. In addition spectral reflectance can be used for forecasting used restoration treatments, natural fading processes, and so on. This study introduces a technique with the use of spectral reflectance to investigate authentic historical materials. In this paper, spectral reflectance is used to detect metameric areas in the materials. Metamerism is a well known phenomenon; two color stimuli are called metamers if they have the same tristimulus values but different spectral radiant power distributions. In the conservation and restoration of historical materials, evidence of previous restoration treatments provides important information. In such restoration treatments, colorants having metamerism relation to the original colorants could been used. Thus, metameric area could tell us which parts of an object have been possibility repainted. The spectral information of the historical materials is necessary to investigate the metamerism in the materials. We have developed a spectral imaging system enabling the direct measurement of the spectral reflectance. The system was used for measurements of ten icons stored in University of Joensuu, Finland. All Icons has been painted in the middle of the 19th century on wooden plate by using natural pigments. Each icon was placed on the sample holder perpendicular to the spectral camera. The light source was a simulated daylight source, and the icons were illuminated in 45-degree angle to their surface. The measured spectral reflectance images were used for the investigation of metamerism. The CIE metamerism index shows the degree of metamerism of two objects having different spectra that match in color under a given illuminant. The CIE metamerism index is calculated in terms of the color difference observed between two objects but under another given illuminant. In this study, this concept is expanded to the spectral domain: the metameric areas in each icon image are detected by using a metamerism coefficient introduced in this study. The metamerism coefficient is calculated from color and spectral information between the reference pixel and test pixel in the spectral reflectance image. In the technique, a reference pixel is first set, and then metamers to the reference pixel are searched by pixel-wise procedure. The performance of the proposed technique is confirmed by using a metamerism test chart, and applied to the ten authentic icons. The technique demonstrates sufficient performance for the test chart, however methods with scientific or chemical analyses such as use of the X-ray diffractometor are required to conclude whether the detected metamers are correct for the icons. As future works,
140
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6062: Spectral Imaging: Eighth International Symposium on Multispectral Color Science
surface reflection property and color mixture model are necessary to enhance accuracy of the detection because the test chart has a flat surface with homogenous mixture of pigments, while the icons have uneven surfaces with complex structure of color pigments and their mixtures.
6062-24, Session 7 Real-rime, multispectral, color acquisition and display using commodity hardware components
6062-22, Session 6 A scanning device for multispectral imaging of paintings C. Bonifazzi, Univ. degli Studi di Ferrara (Italy); P. Carcagnм, A. D. Patria, Istituto Nazionale di Ottica Applicata (Italy); S. Ferriani, ENEA (Italy); R. Fontana, M. Greco, M. G. Mastroianni, M. Materazzi, E. M. Pampaloni, A. Romano, Istituto Nazionale di Ottica Applicata (Italy) A scanning device for 32-band multi-spectral imaging of paintings in the 380ч800 nm spectral region is presented. The system is based on contact-less and single-point measurement of the spectral reflectance factor. Multi-spectral images are obtained by scanning the painted surface under investigation. An adjustment procedure was established and calibration was performed by means of a set of seven matt ceramic color tiles certified by National Physical Laboratory (UK). Colorimetric calculations were carried out in the XYZ colorimetric space, by following the CIE recommendations and choosing the D65 standard illuminant and the 1931 standard observer. Measurement campaigns were carried out on several paintings in situ and at the INOA Optical Metrology Laboratory located inside the Opificio delle Pietre Dure in Florence. As an example we report herein on the measurements carried out on the Madonna in gloria tra Santi by Andrea Mantegna, at present in the Pinacoteque of the Castello Sforzesco in Milan. Multivariate image analyses (MIA) were performed by considering the multi-spectral images as three-way data set. The stack of detected images were unfolded in a 2D data matrix and analyzed by the conventional Principal Component Analysis (PCA). 6062-23, Session 7 Spectral video intraframe compression and database J. P. Purmonen, M. Hauta-Kasari, J. Tuomela, Joensuu Yliopisto (Finland); M. Yamaguchi, M. Mitsui, Tokyo Institute of Technology (Japan); H. Fukuda, National Institute of Information and Communications Technology (Japan) The multiband video camera and display system has developed to increace quality of the video image. The multiband means that there has more than three color channels. The new applications can be created by increasing the number of camera primaries, for example an influence of illumination can be fixed accurately. Furthermore, increasing number of primaries gives a good possibility for image and video processing. However, the growed amount of data causes problems for data transform, store, and process. In this paper, we introduce six band video capturing system and our spectral video database. In the database there is a lot of different type of video clips and the size of spectral video database is more than 300 giga bytes (GB). We also have developed compression scheme for spectral video based on principal component analysis (PCA) and JPEG2000 methods. Here, we concentrate to compress spectral video sequence frame by frame.
D. L. Lau, A. M. Tan, Univ. of Kentucky The focus of this paper is on the broad extension of multi-spectral color to both scientist and consumer by creating camera/projector arrays composed of commodity hardware. In contrast to expensive, high-maintainance systems which rely on the physical registration of device spaces, we rely on the virtual alignment of viewing spaces in software where real-time alignment is achieved using the processing capacity of the graphical processing units of consumer PC video cards. Specifically, this paper focuses on the inclusion of real-time, composite pattern, structured light illumination (SLI) as a means of recording the 3D shape of objects, which will then be used for the registration of single-color images taken from multiple view points simultaneously. As such, the described system is able to achieve a cost per unit that scales linearly with the number of color primaries. 6062-25, Poster Session Construction of multichannel camera gamuts S. Helling, RWTH Aachen (Germany) Device gamuts are commonly defined for output devices, such as monitors or printers. In this paper, a definiton of gamuts of input devices will be examined, considering multispectral cameras as example. A method appropriate to calculate them as a function of the camera model and the spectral reconstruction algorithm will be proposed. The method will be applied to multispectral camera models with a variable number of channels. The characteristics of the resulting gamuts will be shown and examined as a function of the number of channels. Implications on the minumum number of channels needed will be derived. The method proposed here to characterize input devices can be used in addition to common quality criteria such as color distances like CIEDE00, spectral errors, etc. The advantage of the proposed method is the independence of any given spectral data set. This makes it a quality criterion universal for linear (multispectral) cameras and reconstruction algorithms. 6062-27, Poster Session Importance of the texture features in a query from spectral image databases O. Kohonen, M. Hauta-Kasari, Univ. of Joensuu (Finland) A new, semantically meaningful technique for querying the images from a spectral image database is proposed. The technique is based on the use of both color- and texture features. The color features are calculated from spectral images by using the Self-Organizing Map (SOM) when methods of Gray Level Co-occurrence Matrix (GLCM) and Local Binary Pattern (LBP) are used for constructing the texture features. The importance of texture features in a querying is seen in experimental results, which are given by using a real spectral image database. Also the differences between the results gained by the use of co-occurrence matrix and LBP are introduced. It is shown that the texture features are quite a powerful addition to the earlier proposed searching techniques. In our tests the LBPmethod seemed to work better than the co-occurrence matrix. However, there are still some things which have to be considered when combining the texture features together with the color features. The weighting of the different kind of features is one of those things as well as the case-specificity of the query images.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
141
Conf. 6063: Real-Time Image Processing III Monday-Tuesday 16-17 January 2006 Part of Proceedings of SPIE Vol. 6063 Real-Time Image Processing 2006
6063-01, Session 1 Fast computation of free-form deformations in 3D images using FPGAs C. R. Castro-Pareja, Intel Corp.; R. Shekhar, Univ. of Maryland/ Baltimore Free-form deformations, which use B-splines to model a threedimensional curved space, are commonly used to model local deformation fields in elastic image registration algorithms. Fast computation of 3D deformation fields is critical to bringing the application of automated elastic image registration algorithms to routine clinical practice. However, it lies beyond the computational power of current microprocessors, therefore requiring implementations using either massively parallel computers or application-specific hardware accelerators. The use of massively parallel computers in a clinical setting is not practical or costeffective, therefore making the use of hardware accelerators necessary. We present a hardware pipeline that allows accelerating the computation of 3D deformation fields to speeds up to two orders of magnitude faster than software implementations on current workstations and about 30 times faster than other previously reported architectures. The pipeline implements a version of the free-form deformation calculation algorithm, which is optimized to minimize the number of arithmetic operations required to calculate the transformation of a given set of neighboring voxels, thereby achieving an efficient and compact implementation in hardware which allows its use as part of a larger system. The B-spline control points are stored in internal, on-chip memory, thereby allowing fast, parallel accesses. 6063-02, Session 1 Toward real-time stereoscopic depth reconstruction in laparoscopic surgery B. J. McCullagh, F. P. Shevlin, The Univ. of Dublin, Trinity College (Ireland) In order to achieve near real-time frame rates, stereoscopic depth reconstruction applications have to use local methods, such as sum of absolute difference, rather than global methods, such as graph cuts. In this paper we propose some novel methods which will speed up correlation based methods so that more complex algorithms and disparity map refinement can be applied without reducing the frame rate. Encoding a series of images as an MPEG-2 video sequence produces motion vectors relating 16x16 pixel regions of each frame with the previous frame. Encoding a stereo pair as an MPEG sequence provides a series of estimates which can narrow the search space for correlation type algorithms, decreasing the computation time. Extracting the motion vectors from the left or right sequence reveals areas of the scene that have remained unchanged, and can therefore retain the same disparity value. This is especially useful for applications where robotic arms are moving against an almost static background (laparoscopic surgery etc). Programmable GPUs have attracted much attention recently and the implementation of some local methods and comparison of execution speed between CPU and GPU are discussed along with the decrease in computation time achieved using motion vectors. 6063-03, Session 1 Real-time wavelet denoising with edge enhancement for medical x-ray imaging G. Luo, D. Osypiw, Buckinghamshire Chilterns Univ. College (United Kingdom)
X-ray image visualized in real-time plays an important role in clinical applications. The real-time system design requires that images with the highest perceptual quality be acquired while minimizing the x-ray dose to the patient, which can result in severe noise that must be reduced. The approach based on the wavelet transform has been widely used for noise reduction. However, by removing noise, high frequency components belonging to edges that hold important structural information of an image are also removed, which leads to blurring the features. This paper presents a new method of x-ray image denoising based on fast lifting wavelet thresholding for general noise reduction and spatial filtering for further denoising by using a derivative model to preserve edges. General denoising is achieved by estimating the level of the contaminating noise and employing an adaptive thresholding scheme with variance analysis. The soft thresholding scheme is to remove the overall noise including that attached to edges. A new edge identification method of using approximation of spatial gradient at each pixel location is developed together with a spatial filter to smooth noise in the homogeneous areas but preserve important structures. Fine noise reduction is only applied to the non-edge parts, such that edges are preserved and enhanced. Experimental results demonstrate that the method performs well both visually and in terms of quantitative performance measures for clinical x-ray images contaminated by natural and artificial noise. The proposed algorithm with fast computation and low complexity provides a potential solution for real-time applications. 6063-05, Session 2 Real-time high-level video understanding using data warehouse B. Lienard, X. Desurmont, B. Barrie, J. Delaigle, Multitel A.S.B.L. (Belgium) High-level Video content analysis such as video-surveillance is often limited by computational aspects of automatic image understanding, i.e. it requires huge computing resources for reasoning processes like categorization and huge amount of data to represent knowledge of objects, scenarios and other models. This article explains how to design and develop a "near real-time adaptive image datamart", used, in a first time as mass storage, and in a second time as a decisional support system for vision algorithms. Using RDF specification to store data coming from vision algorithms, we can optimize the datawarehouse mechanism and add some processes able to adapt the current model and pre-process data to speed-up queries. In this way, when new data are sent to the datamart for long term storage, using object-oriented interfaces to simplified queries and given distributed computing feature, it is processed and in memory data-model is updated. After some processing possible interpretations of these data is returned back to the caller. To demonstrate this new approach, we present typical scenarios applied to this architecture such as people tracking and events detection in a multi-camera network and we will show how this system becomes a high-semantic data container for external datamining. 6063-06, Session 2 Video surveillance using distance maps T. E. Schouten, H. C. Kuppens, Radboud Univ. Nijmegen (Netherlands); E. L. van den Broek, Vrije Univ. Amsterdam (Netherlands) and Katholieke Univ./Nijmegen (Netherlands) A distance transformation generates a distance map image in which the value of each pixel is its distance to a given set (O) of pixels. Different kind of surveillance parameters can be derived from the
142
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6063: Real-Time Image Processing III
distance maps. For example, lack of maneuvering space for moving objects can be noticed and, hence, risks on collisions can be minimized. However, such applications are only useful when realtime video processing can be utilized. Exact ED maps cannot be obtained using local distances alone. Therefore, we have developed a Fast Exact ED (FEED) transformation where each pixel in O feeds its EDs to all pixels, which in turn take the minimum of all received EDs. Using several speed-up methods FEED proved to be the fastest exact Euclidean distance mapping currently available. For a sequence of frames where only some of the objects move, ED maps can be obtained even faster by saving suitable information from frame to frame. Further methods have been developed for fast detection of fixed and moving objects and for determining the distance of each moving object to the other objects. For each controllable moving object its distance map to the other moving objects and to the fixed objects is determined. 6063-07, Session 2 Vehicle traffic video data real-time processing M. Andreae, W. K. Cheng, Massachusetts Institute of Technology Modified Particle Image Velocimetry (PIV) algorithms offer a computationally efficient method of extracting automobile traffic information from video images. In order to apply the PIV algorithms, the original image is reduced to a sparse image based on motion and edge detection. Then the PIV algorithms extract the velocity vector field from consecutive sparse images. These vectors are then segmented to identify individual vehicles. The speed of the algorithms allows this to be accomplished in real-time at normal video framing rate, thus eliminating the need to store or transmit large quantities of data for post-processing. To test the algorithms, data was collected from a busy 4-way urban intersection. The algorithms were able to extract from the images the number of vehicles, the direction of travel, the size of the vehicles, and the path the vehicles traveled over time in real time (at 15 frames per second). The PIV based algorithms were compared to a segmentation-based set of algorithms and were found to be much faster. Image quality was however found to have a significant impact on the algorithm accuracy. 6063-08, Session 2 Vehicle counting system using real-time video processing P. Crisostomo-Romero, Pontificia Univ. Catolica del Peru (Peru) Transit studies are of great importance for planning a road network with optimal vehicular flow. A vehicular count is essential. There are many methods to perform it, with their respective advantages and disadvantages in installation, reliability and cost. This article presents a vehicle counting system based on video processing. An advantage of such system is the greater detail that is possible to obtain in the shape, size and speed of vehicles. The presented system uses a video camera placed 6 meters above the street level. Fast image processing algorithms and small image dimensions are used to allow real-time processing. Digital filters, mathematical morphology, segmentation and other techniques are used to identify and count all vehicles in the image sequences. The system was implemented under Linux in a 1.8 GHz Pentium 4 computer. It obtains a successful count with frame rates of 15 frames per second for images of size 240x180 pixels and 24 frames per second for images of size 180x120 pixels, thus being able to count all the vehicles whose speeds do not exceed 150 km/h.
6063-09, Session 3 Real-time auto white balancing using DWTbased multiscale clustering N. Kehtarnavaz, N. Kim, M. N. Gamadia, The Univ. of Texas at Dallas Auto white balancing (AWB) involves the process of making white colors to appear as white under different illuminants in digital imaging products such as digital still cameras. This paper presents a computationally efficient auto white balancing algorithm for real-time deployment in imaging products. The algorithm utilizes DWT (discrete wavelet transform) to perform multi-scale clustering (MSC), thus generating a computationally efficient implementation of the original MSC algorithm. The paper also discusses the steps taken to allow running this algorithm in real-time on a digital camera processor. The results of an actual implementation on the Texas Instruments TMS320DM320 processor are provided to illustrate the effectiveness of this algorithm in identifying an illuminant as compared to the widely used gray-world auto white balancing algorithm. 6063-10, Session 3 Real-time antialiasing using adaptive directional filtering P. Rokita, Politechnika Warszawska (Poland) In this paper we present an enhanced real-time selective antialiasing solution. We propose to use a directional filtering technique as an antialiasing tool. The best post-processing antialiasing effect will be obtained if we apply the lowpass filter along local orientation of antialiased features. Previously authors proposed a complicated curve fitting method as a solution for the local feature antialiasing. Here we propose a more simple and efficient solution. Instead of using a curve fitting method based on second order intensity derivatives, we can use directly a set of first order derivatives applied on the z-buffer content. For each feature direction detected an appropriate directional Gaussian convolution filter can be applied. This way the lowpass filter is applied along local features selected for antialiasing, filtering out high frequency distortions due to intermodulation. In this approach the highpass convolution filtering applied on the z-buffer has a twofold application: it selects the objects edges that need to be antialiased and it gives a local feature direction allowing for edge reconstruction. The advantage of the approach proposed here is that it preserves texture details. Textures usually are filtered independently using trilinear or anisotropic filtering, which with traditional antialiasing techniques leads to overblurring. 6063-11, Session 3 A fast eye detector using corners, color, and edges L. Chen, C. Grecos, Loughborough Univ. (United Kingdom) Eye detection plays a central role in an automatic face detection systems and it is also important for human-computer interaction and face tracking. In this paper, we present a novel, unsupervised scheme for detecting eyes in skin patches based on our previous work on skin patch detection. Working on the normalized RGB color space, we use a combination of corner identification, color and edges as heuristics for eyes detection. The proposed scheme consists of four major steps. In the first step, a Harris corner detector is used to detect some points of interest in the skin patch. The second step is the detection of local gray pixels to build up a gray map. The third step is to build up an edge map of the image. The final step of our scheme is combining the Harris corners detected, edge map and the gray map to filter out useless corners,
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
143
Conf. 6063: Real-Time Image Processing III
and finally locate the eyes. Experimental results show that our scheme is very fast in the AR and Champion databases, while retaining very high detection rates. 6063-12, Session 3 Real-time construction of covariance matrices for arbitrary size image windows F. M. Porikli, O. Tuzel, Mitsubishi Electric Research Labs. We present a fast, integral image based algorithm to compute feature covariance matrices within all arbitrary size rectangular regions in an image. This technique significantly improves the computational load of covariance matrix extraction process by taking advantage of the spatial arrangement of points. Covariance is an essential measure of how much the deviation of two or more variables or processes match. In our case, these variables correspond to point features such as coordinate, color, gradient, orientation, and filter responses. Integral images are intermediate image representations used for calculation of region sums. Each point of the integral image is a summation of all the points inside the rectangle bounded by the upper left corner of the image and the point of interest. Using this representation, any rectangular region sum can be computed in constant time. We follow a similar idea for fast calculation of region covariance. We construct integral images for all separate features as well as integral images of the multiplication of any two feature combinations. Using these set of integral images and region corner point coordinates, we directly extract the covariance matrix coefficients. We show that the proposed integral image based method decreases the computational load to quadratic time. 6063-13, Session 4 High-performance VLSI architecture for adaptive scaling P. P. Dang, STMicroelectronics Scaling is one of the basic operations in image and video processing. It has been used in many consumer products such as printers, set-top boxes, flat-panel displays and high definition televisions (HDTV). In these applications, scaling algorithms are used to enlarge the images, which produce outputs for printers or display devices. This approach saves substantial amount of CPU time since other kernels in the application do not have to process large size input image. Scaling algorithms, in general, are based on interpolation operation. Most of the classical scaling algorithms simply apply a single interpolation technique across the whole image. The outputs from this approach suffer from the artifacts of the blurring of edges, the ringing around the edges, loss of texture or details. In this paper, we introduce an adaptive approach for image scaling. In addition, we present an efficient VLSI architecture to implement the proposed adaptive algorithm in hardware. The proposed algorithm is designed to reduce the artifacts, maintain the sharpness, and produce high-resolution outputs. Experimental results show that it works well for both image and video applications. In this paper, an efficient VLSI architecture was designed to address the real-time constrain for the consumer products. 6063-14, Session 4 Architecture for hardware driven image inspection based on FPGAs J. Fuertler, J. Brodersen, Austrian Research Ctrs. GmbH Seibersdorf (Austria); P. Roessler, Technische Univ. Wien (Austria); K. J. Mayer, Austrian Research Ctrs. GmbH - Seibersdorf (Austria); G. Cadek, C. Eckel, H. Nachtnebel, Technische Univ. Wien (Austria)
Requirements for contemporary print inspection systems for industrial applications include, among others, examination of fine details of the print, inspection from various perspectives and with different spectral sensitivity (visible, infrared, ultraviolet), as well as high throughput. Therefore, an optical inspection system for such tasks has to be equipped with several high-speed/high-resolution cameras, each acquiring hundreds of megabytes of data per second. This paper presents an inspection system which meets the given requirements by exploiting data parallelism and algorithmic parallelism. This is achieved by using complex field-programmable gate arrays (FPGA) for image processing. The scalable system consists of several processing modules, each representing a pair of a FPGA and a digital signal processor. The main chapters of the paper focus on the functionality implemented in the FPGA. The image processing algorithms include flat-field correction, lens distortion correction, image pyramid generation, neighborhood operations, a programmable arithmetic unit, and a geometry unit. Due to shortage of on-chip memory, we use a multi-port memory concept for buffering streams of image data between off-chip and on-chip memories. Furthermore, performance measurements of the actual processing module and the whole system are presented. The paper concludes with an outlook to future projects. 6063-15, Session 4 Using a field programmable object array (FPOA) to accelerate image processing S. Riley, P. Chiang, MathStar, Inc. Digital signal processing has been widely used in the satellite, radar and other surveillance applications. The steadily growing size of the input data array and the sophisticated processing algorithms have continuously added more demand on processing power. The current solutions to these high power computing applications include fixed function ASICs or programmable FPGAs. The ASIC approach can deliver increased computing power but has a high cost, long development time and limited programmability. An FPGA offers the flexibility of modification and upgrades, but individual FPGAs often cannot deliver the required performance. This paper describes a system solution built around a Field Programmable Object Array (FPOA). The FPOA is a reconfigurable architecture that delivers the required high performance without sacrificing programmability. Example FPOA designs for various parts of a satellite image processing engine are described in this presentation. The performance of a complete digital signal processing space satellite application is provided to conclude this paper. 6063-16, Session 4 Novel windowing technique realized in FPGA for radar system E. Escamilla-Hernбndez, Instituto Politйcnico Nacional (Mexico); V. F. Kravchenko, Instituto Politйcnico Nacional (Mexico) and Institute of Radio Engineering and Electronics (Russia); V. I. Ponomaryov, Instituto Politйcnico Nacional (Mexico) To improve the weak target detection ability in radar applications a pulse compression is usually used that in the case linear FM modulation can improve the SNR. One drawback in here is that it can add the range side-lobes in reflectivity measurements. Using weighting window processing in time domain it is possible to decrease significantly the side-lobe level (SLL) and resolve small or low power targets those are masked by powerful ones. There are usually used classical windows such as Hamming, Hanning, etc. in window processing. Additionally to classical ones in this paper we also use a novel class of windows based on atomic functions (AF) theory. For comparison of simulation and experimental results we applied the standard parameters, such as coefficient of amplification, maximum level of side-lobe, width of main lobe, etc.
144
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6063: Real-Time Image Processing III
To implement the compression-windowing model on hardware level it has been employed FPGA. This work aims at demonstrating a reasonably flexible implementation of FM-linear signal, pulse compression and windowing employing FPGA's. Classical and novel AF window technique has been investigated to reduce the SLL taking into account the noise influence and increasing the detection ability of the small or weak targets in the imaging radar. Paper presents the experimental hardware results of windowing in pulse compression radar resolving several targets for rectangular, Hamming, Kaiser-Bessel, Up(x), Fup1(x), sigma(x), fup4,2(x)B2(x), Fup4(x)D3(x), Fup4(x)D3.5(x), Fup6,2(x)G2,2(x), Fup6(x)G3(x), Fup6,2(x)G3(x) functions windows. The windows created by use the atomic functions offer sufficiently better decreasing of the SLL in case of noise presence and when we move away of the main lobe in comparison with classical windows. 6063-17, Session 4 Real-time hardware for a new 3D display M. Akil, B. Kaufmann, Йcole Supйrieure d'Ingйnieurs en Йlectronique et Йlectrotechnique (France) We are working on the study and the design of a new 3D display system based upon auto stereoscopy. To control and make the 3D device working, we need to generate appropriate images. These images will be generated from 3D data sent by a computer to the 3D image processing chain. This processing chain is composed by: 3D geometry, rasterization and voxels processing. In order to reach a speed suitable for real-time display (up to 30x3D frames per second projected to 200 different directions, which means 6,000 projected images per second), the use of a dedicated hardware is necessary. In this article, we focus our presentation on the voxels processing. For this step, we propose a new coding algorithm. We propose a dedicated hardware to implement efficiently this coding algorithm, because this is critical part of our 3D processing chain. The implementation uses chained lists to store depths and patterns. The voxels inside patterns are coded as a chained list to fasten the coding. We obtain a real time processing which is able to store data at 10MBx30 images per second. In this article we describe a dedicated architecture of the processing chain of our new 3D display. We also simulated and validated the coding hardware and the overall architecture chain by using system C and Visual Elite software tool. 6063-18, Session 4 A rapid prototyping methodology to implement and optimize image processing algorithms for FPGAs M. Akil, P. Niang, T. Grandpierre, Йcole Supйrieure d'Ingйnieurs en Йlectronique et Йlectrotechnique (France) In this paper we present a seamless flow of transformations that leads to the generation of a complete VHDL design corresponding to the optimized implementation of image processing algorithms specified by Factorized and Conditioned Data Dependence (FCDD) graph model. The aim of our design flow based upon AAA/SynDEx extension is to implement a real-time application onto specific integrated circuits. This extension uses a Factorized (to specify Loop) and Conditioned (to specify if then else structure) Data dependence graph model, from the algorithm specification down to the architecture implementation, through optimization expressed in terms of defactorization (unrolling) transformations. This paper presents the transformation flow used by our methodology to automate the hardware implementation process of image processing algorithms onto reconfigurable circuits. We illustrate the proposed design flow for the hardware implementation on the Spartan XC2S100 Xilinx board of the image processing filters, edge detection operators (operators based on spatial convolution, Canny and Deriche algorithms), and DCT (cosine transform) and DCT-1 .
6063-19, Poster Session An efficient illuminance-reflectance nonlinear video stream enhancement model L. Tao, V. K. Asari, Old Dominion Univ. An efficient real-time video stream enhancement algorithm based on illuminance-reflectance model is proposed for improving the visual quality of digital video streams captured under insufficient and/or non-uniform lighting conditions. The paper presents computational methods for estimation of scene illuminance and reflectance, adaptive dynamic range compression of illuminance, and adaptive enhancement for mid-tone frequency components. Adaptive illuminance enhancement is a global intensity transformation based on a Windowed Inverse Sigmoid (WIS) nonlinear transfer function, the curve shape of WIS is determined by its window range which is self-tuned by the histogram statistics of the input image. This process largely increases the luminance of darker pixels and compresses the dynamic range of the image at the same time. Due to the degradation of mid-tone frequency features caused by illuminance dynamic range compression, a power-function based adaptive mid-tone frequency enhancement is utilized as a compensation by tuning the frequency components of the image (in frequency domain) based on the magnitude of each pixel in spatial domain with respect to its neighboring pixel values, and this process is adaptively controlled by the global statistics of the image. This algorithm is an effective technique for image enhancement with simple computational procedures, which makes real-time enhancement for homeland security application successfully realized. It also demonstrates its robust and high quality performance while enhancing video streams. 6063-21, Poster Session A novel two-pass hexagonal search algorithm for motion estimation Y. Wu, The Univ. of Reading (United Kingdom) This paper presents a novel two-pass algorithm constituted by Linear Hashtable Motion Estimation Algorithm (LHMEA) and Hexagonal Search (HEXBS) for block base motion compensation. On the basis of research from previous algorithms, especially an onthe-edge motion estimation algorithm called hexagonal search (HEXBS), we propose the LHMEA and the Two-Pass Algorithm (TPA). We introduce hashtable into video compression. In this paper we employ LHMEA for the first-pass search in all the Macroblocks (MB) in the picture. Motion Vectors (MV) are then generated from the firstpass and are used as predictors for second-pass HEXBS motion estimation, which only searches a small number of MBs. The evaluation of the algorithm considers the three important metrics being time, compression rate and PSNR. The performance of the algorithm is evaluated by using standard video sequences and the results are compared to current algorithms. Experimental results show that the proposed algorithm can offer the same compression rate as the Full Search. LHMEA with TPA has significant improvement on HEXBS and shows a direction for improving other fast motion estimation algorithms, for example Diamond Search. 6063-22, Poster Session Real-time image processing based on robust linear combinations of order statistics F. J. Gallegos-Funes, J. L. Varela-Benitez, V. I. Ponomaryov, Instituto Politйcnico Nacional (Mexico) In this paper we present the capability and real-time processing features of a new type of L-filter for the removal of impulsive and multiplicative noise in real-time image processing applications. The
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
145
Conf. 6063: Real-Time Image Processing III
proposed filter uses the robust RM-estimator in the filtering scheme of L-filter according with the RM-KNN filtering algorithms. Extensive simulation results in known reference images, and some medical and SAR images have demonstrated that the proposed filter consistently outperforms other filters by balancing the tradeoff between noise suppression and detail preservation. The criteria used to compare the performance or various filters were the PSNR, MAE, and processing time. The real-time implementation of proposed algorithm was realized on the DSP TMS320C6701. The processing time of proposed filter includes the time of data acquisition, processing and store data. We found that the processing time values of proposed filter depend of the image to process and do not practically vary for different complex noise level; these values depend also of the calculation of influence functions, parameters of the proposed filter, and different distribution functions used to calculate the coefficients of the new type of L-filter. 6063-23, Poster Session A new concept of real-time security camera monitoring with privacy protection by masking moving objects K. Yabuta, H. Kitazawa, T. Tanaka, Tokyo Univ. of Agriculture and Technology (Japan) We present a novel framework for real-time encoding of images obtained by a security monitoring camera with protecting the privacy of moving objects in recorded images. We are motivated by the fact that although security monitoring cameras can deter crimes, they may infringe the privacy of those who and objects which are monitored by the cameras. Two methods are proposed to make moving objects unrecognizable. One is the scrambling and the other is erasing in which moving objects are replaced by the corresponding background images. In order to reconstruct the original image, the moving object images are kept in a JPEG bitstream by using watermarking with encryption by the advanced encryption standard (AES). Therefore, a normal user using a JPEG viewer can only see masked images, where the moving objects are unrecognizable or completely invisible. Only a special viewer with a keyword can reconstruct the recorded images with the original objects. In the case where there are multiple moving objects, only a desired object can be decoded by using the moving object tracking information, while the other objects are kept unrecognizable. Real-time processing was achieved by using distributed computing. The frame rate was nearly 10 frames/sec. 6063-24, Poster Session Online monitoring for wood pieces on a moving conveyor belt W. Wang, Chongqing Univ. of Posts and Telecommunications (China) To make automation in different industrial applications, vision based system development is a kind of computer systems for autodetection and monitoring of online input material and output products. In combustion industry or wood production industry, one kind of burning materials is waste wood pieces. This paper presents a Windows based system for image analysis and computer vision of wood piece materials. The system was designed for auto-detection of wood piece materials on a moving conveyor belt, and developed on Microsoft Windows on PC computer. To make the system work efficiently in the plant, we designed and developed a flexible Windows platform which mainly consists of image acquisition, image processing, wood piece delineation and analysis, and interface between different computers in use. Hundreds of functions for image analysis and statistics are included in the system, especially in the
wood piece delineation part. A number of newly-developed algorithms can delineate wood pieces with high accuracy and high speed, and in the wood piece analysis part, each wood piece can be characterized by 50-60 parameters that can also be used for constructing wood piece models directly in a Windows platform environment. The system makes online measurement easy. An online database is built in the system, making old measurement data easily checkable. 6063-25, Poster Session A hardware-accelerated approach to computing multiple image similarity measures from joint histogram C. R. Castro-Pareja, Intel Corp.; R. Shekhar, Univ. of Maryland/ Baltimore A hardware-accelerated approach to computing multiple image similarity measures from joint histogram Image similarity-based image registration is an iterative process that, depending on the number of degrees of freedom in the underlying transformation, may require hundreds to tens of thousands of image similarity computations to converge on a solution. Computation time often limits the use of such algorithms in real-life applications. We have previously shown that hardware acceleration can significantly reduce the time required to register two images. However, the hardware architectures we presented previously were limited to mutual information calculation, which is one of several commonly used image similarity measures. In this presentation, we will show how our architecture can be adapted for the calculation of other image similarity measures such as cross-correlation and rms difference in approximately the same time and using the same hardware resources as those for the mutual information case. As in the case of mutual information calculation, the joint histogram is calculated as a first step. Both cross-correlation and rms difference can be calculated with a single pass through the joint histogram, which normally takes only a small fraction (~1%) of the time required to calculate the joint histogram itself. 6063-26, Poster Session Real-time human detection by shape and motion H. Ran, Wuhan Univ. of Technology (China) In this article, we concentrate on visible-spectrum vision systems for human figure detection that could be successfully implemented in cars for driver support in urban scenes or for autonomous driving. We present here a new classification based on the kind of approaches used in the first place to detect a candidate for a pedestrian, and secondly to recognise a person among the collected candidates and elimination of false positives by its motion. A first distinction among different methods is whether they distinguish two steps in the process (a true detection step and a recognition step) or do not. We first search the entire image for a pedestrian, most often relying on a shape based and multi-scale analysis for the upper body contour. The authors here report a high effectiveness in unpredictable environments, where background segmentation is complicated by uncontrollable variations of the scene. Preliminary experiments on this dataset of 6 videos showed detection rates ranging from 75%-90% with one false alarm per frame on an average. The implementation is quite robust to noise and occlusions.
146
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6063: Real-Time Image Processing III 6063-31, Poster Session Uranus: an environment for rapid prototyping of real-time video processing based on FPGA M. A. Nuсo-Maganda, V. H. Rosales-Hernбndez, L. N. CastilloJimenez, G. Sosa-Ramнrez, M. O. Arias-Estrada, Instituto Nacional de Astrofнsica, Уptica y Electrуnica (Mexico) In this work a tool called URANUS for video processing algorithms prototyping is proposed. URANUS takes as input a set of image processing processes, coded in C or Handel-C Languages, and generates the equivalent set of processes mapped into a specific FPGA device. URANUS makes possible the data interchange between PC and FPGA by PCI, Ethernet, USB, Serial or Parallel Ports. The configuration parameters of each FPGA platforms are defined in URANUS. However, the interaction between PC and FPGA is transparent for the users, since the user only chains a set of a desired image processing algorithms. The implemented processes can have many types of input and output information. A high level of abstraction is used to generalize the algorithm implementation for making easy the Software to Hardware migration. A set of Hardware and Software libraries were developed, and final users can add their own processes following some guidelines. URANUS generates reports analysing the source code to migrate it to a specific hardware platform, such as execution time, time consumed in each function and number of operations. The platforms actually supported by URANUS are: Alphadata ADM-XPL and Celoxica RC200. An example of target tracking application is documented in this work. 6063-32, Poster Session Determination of traffic intensity from camera images using image processing and pattern recognition techniques M. Mehrubeoglu, Texas A&M Univ./Corpus Christi; L. McLauchlan, Texas A&M Univ./Kingsville This paper investigates traffic intensity from webcam images. Images are downloaded from live traffic webcams from various highways or intersections at various intervals. Traffic intensity is investigated at different times of the day (currently daylight hours only) under various weather conditions. Images are transformed into binary images using an adaptive threshold value based on the mean intensity of the region of interest (ROI) in each image. Areas outside ROI that are not part of the road and vehicles on the road are masked and excluded from processing to reduce processing time. Horizontal edges could be obtained from vehicles that corresponded to the intensity transitions across different parts of the vehicles and the road. Horizontal directional patterns were extracted using a 3X3 Prewitt's window. Extracted patterns (horizontal edges that were of predetermined length and width) were recognized using template matching. Once a single pattern was identified, a scaled area based on perspective vision around the location of the found pattern, was excluded from search for the next pattern to avoid detecting multiple patterns from the same vehicle. Each matched pattern was considered to correspond to a vehicle. Traffic intensity was estimated from the number of vehicles identified by the algorithm.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
147
Conf. 6064A: Image Processing: Algorithms and Systems V
Monday-Tuesday 16-17 January 2006
Part of Proceedings of SPIE Vol. 6064 Image Processing: Algorithms and Systems,
Neural Networks and Machine Learning 6064A-01, Session 1
6064A-04, Session 1
Affine invariant surface evolutions for 3D image segmentation
Edge-based stochastic active contours for medical imaging
Y. Rathi, Georgia Institute of Technology; P. Olver, G. Sapiro, Univ. of Minnesota; A. R. Tannenbaum, Georgia Institute of Technology In this paper we present an algorithm for 3D medical image segmentation based on an affine invariant flow. The algorithm is simple to implement and semi-automatic. The technique is based on active contours evolving in time according to intrinsic geometric measures of the image. The surface flow is obtained by minimizing a global energy with respect to an affine invariant metric. Affine invariant edge detectors for 3-dimensional objects are also computed which have the same qualitative behavior as the Euclidean edge detectors. Results on artificial and real MRI images show that the algorithm performs well, both in terms of accuracy and robustness to noise. 6064A-02, Session 1 Iterative Markovian estimation of mass functions in Dempster Shafer evidence theory: application to multisensor image segmentation L. Bentabet, M. Jiang, Bishop's Univ. (Canada)
J. J. Traisnel, A. R. Tannenbaum, Georgia Institute of Technology Active contour methods have proven to be a useful tool in medical image segmentation. Edge based methods while very popular have several problems including sensitivity to noise; indeed noise and secondary edges may act as local minima about which the evolving contour may get stuck. To overcome this issue, researchers have attempted to formulate models which integrate region-based terms and probabilistic models of the image. In this note, we propose a stochastic version of the geometric active contour for the segmentation of noisy medical imagery based on recent work on Stochastic Partial Differential Equations. The idea is to explicitly adding a stochastic perturbation to the evolution equation. The model has some nice advantages including the fact that one does not need to incorporate any prior knowledge of the image in order to segment the relevant structures. We indicate how to implement the introduction of a given stochastic perturbation in the level set framework, and described several noise models. Experimental results are provided on both real medical images and synthetic images and seem to demonstrate that adding a stochastic term improves the standard edge-based active contour methods by overcoming the problem of a contour being stuck at a local minimum.
Mass functions estimation is a key issue in evidence theory-based segmentation of multisensor images. In this paper, we generalize the statistical mixture modeling and the Bayesian inference approach in order to quantify the confidence level in the context of DempsterShafer theory. We demonstrate that our model assigns confidence levels in a relevant manner. Contextual information is integrated using a Markovian field that is adapted to handle compound hypotheses. The multiple sensors are assumed to be corrupted by different noise models. In this case, we show the interest of using a flexible Dirichlet distribution to model the data. The effectiveness of our method is demonstrated on synthetic and radar and SPOT images. 6064A-03, Session 1 Progressive halftoning by Perona-Malik error diffusion and stochastic flipping J. J. Shen, Univ. of Minnesota Halftoning has been a significant topic in image processing due to many emerging applications, various diversified approaches, and challenging theoretical analysis. Inspired by the wealthy literature on halftoning, as well as the recent PDE (partial differential equations) approach in image processing, the current work proposes a novel progressive halftoning algorithm by empolying the celebrated anisotropic diffusion model of Perona and Malik (IEEE Trans. Pattern Anal. Machine Intell., 12:629-639, 1990), and a properly designed stochastic strategy for binary flipping. The halftone outputs from the proposed model are typical samples of some random fields, which share many virtues of existent deterministic halftone algorithms, as well as show many interesting features like the blue noise behavior. The new model is independent of traditional windows, tiles, or paths, and allows direct parallel implementation.
6064A-05, Session 1 Multiple wavelet coherence analysis S. C. Olhede, G. Metikas, Imperial College London (United Kingdom) We propose a method for analysis of localised relationships between multiple images, that naturally treats the local phase structure and orientation of any variation in the observed images. The method is based on several nonseparable wavelet decompositions of the images. The set of mother wavelets is constructed by finding optimally concentrated orthogonal wavelet functions, and each isotropic wavelet is extended to a triplet of functions. The full set of wavelet transforms of two images can be used to extract local oscillatory components of the images present at a given spatial and scale point, and subsequently used to determine the local attenuation and phase shift between the two images. The determination of the local phase and orientation involves calculating the continuous wavelet transform of the images, then forming the scalogram matrix, and calculating the wavelet coherence. Robust estimates of noisy images can be constructed by averaging over wavelet coefficients, extending Thomson's method to isotropic localised two-dimensional analysis, and we demonstrate the reduced variability in the estimators by this approach. 6064A-06, Session 1 New class of interpolation methods based on discretized lie group transforms A. Zaratsyan, J. Patera, Univ. de Montrйal (Canada); H. Zhu, York Univ. (Canada) We propose a new class of interpolation methods for digital images using variants of the discrete group transform (DGT) recently developed by our group. The DGT is a transform of functions given on a finite region of the plane on a rank-2 simple Lie group as their expansions into series of special functions, called the C-functions (in recognition of the fact that the functions generalize cosine to any finite dimensions). The C-functions are orthogonal and readily discretized on lattices of appropriate symmetry and any density. The
148
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6064A: Image Processing: Algorithms and Systems V
traditional two-dimensional discrete cosine transform (DCT) is the simplest particular case of our transforms. Image processing using the DGT has shown successes in the military and medical applications. The inferiority of the DCT interpolation is qualitatively understood as interpolation in only two orthogonal directions, the variants of the DGT defined on the other three rank-2 groups can invoke higher quality interpolations: 4 directions for square lattice (two orthogonal and two diagonal directions), 3 or 6 directions for triangular lattice. Therefore, interpolation through the DGT may substantially improve the quality of the interpolated image. Here, the DGT is conceptually introduced and the potential of the DGT in interpolation is demonstrated, using square lattices for simplicity. Quantitative and qualitative comparisons of the interpolation using variants of the DGTs and some standard methods are presented. Both theoretical, as well as some real-world examples are used in the comparisons. The robustness of performance in the presence of noise is also investigated and their computational complexities are discussed. 6064A-07, Session 1 Optimization procedures for the estimation of phase portrait parameters in orientation fields F. J. Ayres, R. M. Rangayyan, Univ. of Calgary (Canada) Oriented patterns in an image often convey important information regarding the scene or the objects contained. Given an image presenting oriented texture, the orientation field of the image is the map that depicts the orientation angle of the texture at each pixel. Rao and Jain developed a method to describe oriented patterns in an image based on the association between the orientation field of a textured image and the phase portrait generated by a pair of linear first-order differential equations. The optimization of the model parameters is a nonlinear, nonconvex optimization problem, and practical experience shows that irrelevant local minima can lead to convergence to inappropriate results. We investigated the performance of three optimization algorithms for the estimation of the best phase portrait parameters, for a given orientation field. The investigated algorithms were: nonlinear leastsquares, linear least-squares, and simulated annealing. The algorithms are evaluated in terms of the error between the estimated parameters and the parameters known by design, in the presence of noise in the orientation field and imprecision in the initialization of the parameters. The computational effort required by each algorithm is also assessed. 6064A-08, Session 1 Optimized gradient filters for hexagonal matrices T. Shima, S. Saito, Tokyo Institute of Technology (Japan); M. Nakajima, Tokyo Institute of Technology (Japan) and National Institute of Informatics (Japan) In many cases, digital images are represented by square matrices. For instance, digital cameras, displays, and many systems for vision or image processing use square matrices to represent an image. Square matrices are not, however, the only placement method of pixels. Another placement can be found in, for example, the receptors of the human retina, where a hexagonal placement is observed. In the case of square matrices, the distance between diagonal adjacent pixels is not equal to the one between off-diagonal adjacent pixels. Ando introduced consistent gradient filters to cope with this problem. These filters are derived to minimize inconsistency of gradient filters. In contrast, distances between any adjacent pixels
are the same for hexagonal matrices. The principal advantage of using hexagonal matrices is its isotropy from the perspective of pixel arrangement. In this paper, we derive consistent gradient filters for hexagonal matrices following Ando's method to derive consistent gradient filters for square matrices. The resultant hexagonal consistent gradient filters are compared with square consistent ones. The results indicate that hexagonal consistent gradient filters are superior to square ones in consistency, in proportion of consistency to output power, and in localization. 6064A-09, Session 2 Super-fast Fourier transform S. S. Agaian, O. Caglayan, The Univ. of Texas at San Antonio The discrete Fourier transform (DFT) is the most widely used application in the signal and image processing, and the communication systems. The theory of trigonometric series can be dated back to the beginning of 18th century. In 1747, Euler represented the movements of the planets in the form of a trigonometric series, which actually contain what is now called the Fourier series. In 1807, Fourier discovered that a wide class of signals could be generated by summing scaled sine and cosine functions. Still in 21st century, discrete Fourier transform is one of the most widely applied tools in science and technology. The direct calculation of the DFT and the IDFT is computationally intensive and requires complex multiplications and complex additions. Such computational complexity has been quite costly in terms of reduced signal processing speed, increased power consumption, and higher expense. Therefore, it is natural to ask how we can reduce the arithmetical complexity of DFT. One important tool in modern digital signal processing applications that helps to reduce the high computational cost is the Fast Fourier Transform (FFT). In this paper, a new fast Fourier transform algorithm with linear multiplicative complexity is introduced for real or complex input signals. The proposed algorithm not only reduces the number of multiplications in the computation significantly, but also reduces the total number of operations (arithmetic complexity, which is the number of multiplications and additions) compared to the existing methods. 6064A-11, Session 2 A high-speed rotation method for binary document images based on coordinate operation of run data Y. Shima, H. Ohya, Meisei Univ. (Japan) The rotation of an image is one of the fundamental functions in image processing and is applied to document image processing in the office. A method of image rotation based on digital image data has been developed. This paper assumes the binary digital data, and proposes a method which is different from the traditional one based on pixel data. This method can execute a high-speed rotation of binary image based on coordinate data for the start and the end of the run. The image rotation at an arbitary angle can be realized by the real number operation on the run data, which is suited to the personal processor. It is a practically useful method since the processing is fast and less memory capacity is required. In this paper, a discussion is made first on the format of the run data, the number of runs and the data complexity for the binary data. Then the newly devised rotation for the binary image is described. The rotation method is used to perform successively the skew coordinate transformations in the vertical and horizontal directions, to determine the rotated images. Finally, the processing time was examined to demonstrate experimentally the usefulness of the proposed method.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
149
Conf. 6064A: Image Processing: Algorithms and Systems V
6064A-12, Session 2 A hardware implementation of the discrete Pascal transform for image processing T. J. Goodman, M. F. Aburdene, Bucknell Univ. The discrete Pascal transform is a polynomial transform with applications in pattern recognition, digital filtering, and digital image processing. It already has been shown that the Pascal transform matrix can be decomposed into a product of binary matrices. Such a factorization leads to a fast and efficient hardware implementation without the use of multipliers, which consume large amounts of hardware. We recently developed a field-programmable gate array (FPGA) implementation to compute the Pascal transform. Our goal was to demonstrate the computational efficiency of the transform while keeping hardware requirements at a minimum. Images are uploaded into memory from a remote computer prior to processing, and the transform coefficients can be offloaded from the FPGA board for analysis. Design techniques like as-soon-as-possible scheduling and adder sharing allowed us to develop a fast and efficient system. An eight-point, one-dimensional transform completes in 13 clock cycles and requires only four adders. An 8x8 two-dimensional transform completes in 240 cycles and requires only a top-level controller in addition to the one-dimensional transform hardware. Finally, through minor modifications to the controller, the transform operations can be pipelined to achieve 100% utilization of the four adders, allowing one eight-point transform to complete every seven clock cycles. 6064A-13, Session 3 Using clustering for document reconstruction A. Ukovich, A. Zacchigna, G. Ramponi, G. Schoier, Univ. Degli Studi di Trieste (Italy) The reconstruction of shredded or torn documents is a problem which may arise in the forensics and investigative science fields. Image processing techniques give the opportunity for a computeraided reassembly. The visual content of the pieces is automatically extracted from the digitized remnants and it is represented by numerical features. The problem of the reconstruction is NP-hard. Hence, approximate algorithms must be adopted for its solution. In order to keep down the computational complexity, we propose an approach based on a clustering strategy. The pieces originally belonging to the same page are first grouped together; the following search for the matching remnant is conducted within smaller subsets instead of within the whole set of remnants. Three aspects must be taken into consideration for the clustering to be effective: the selection of suitable numerical features, the choice of the number of clusters, and the clustering algorithm used. We use low-level features related to the font format, the page layout, and the ink-background colours. The set of the most discriminating features may vary according to the specific dataset considered. The estimated number of clusters is related to the shredder used (number of remnants/page) and to the amount of blank remnants/ page. 6064A-14, Session 3 Automatic detection and tracking of reappearing targets in forward-looking infrared imagery A. Bal, M. S. Alam, Univ. of South Alabama Target detection and tracking algorithms deal with the recognition of a variety of targets obtained from a multitude of sensor types such as forward-looking infrared (FLIR), synthetic aperture radar and laser
radar. There are many factors that must be accommodated by the detection and tracking algorithm such as target aspect variations caused by rotation of target or sensor operation (zoom, pan, tilt etc.), different weather conditions, cluttered background, and sensor noise. In addition, temporary disappearance and then reappearance in the field-of-view may be encountered during the tracking processes. To accommodate the above mentioned problems, it is important to develop a robust target model to overcome the problem of distinguishing similar target or non-target objects that already existed in the scene. Since the target history is known before target exiting the current frame, training based techniques have been developed using combination of two techniques; tuned basis functions (TBF) and correlation based template matching (TM) technique. The TBFs obtained for the target class are used to detect possible candidate target images. The detected candidate target images are then introduced into the second algorithm, called clutter rejection module, to determine the target reentering frame and location of the target. The performance of the proposed TBF-TM based reappearing target detection and tracking algorithm has been tested using real-world forward looking infrared video sequences. 6064A-16, Session 3 Robust human motion detection via fuzzy set based image understanding Q. Li, J. You, The Hong Kong Polytechnic Univ. (Hong Kong China) This paper presents image understanding approach to monitor human movement and identify the abnormal circumstance by robust motion detection for the care of the elderly in a home-based environment. In contrast to the conventional approaches which apply either a single feature extraction or a fixed object model for motion detection and tracking, we introduce a multiple feature extraction and a fuzzy set based image understanding scheme for robust motion detection. The proposed algorithms include 1) multiple image feature extraction including the fuzzy compactness based detection of interesting points and fuzzy blobs, 2) adaptive image segmentation via multiple features, 3) Hierarchical motion detection, 4) a flexible model of human motion adapted in both rigid and non-rigid conditions, and 5) Fuzzy decision making via multiple features. 6064A-17, Session 3 k-max: segmentation based on selection of max-tree deep nodes A. G. Silva, S. C. Felipussi, G. L. F. Cassol, Univ. do Estado de Santa Catarina (Brazil); R. de Alencar Lotufo, Univ. Estadual de Campinas (Brazil) This work proposes the segmentation of grayscale image from of its hierarchical region based representation. The Max-tree structure has demonstrated to be useful for this purpose, offering a semantic vision of the image, therefore, reducing the number of elements to process in relation to the pixel based representation. In this way, a particular searching in this tree can be used to determine regions of interest with lesser computational effort. A generic application of detection of peaks is proposed through searching nodes to kup steps from leaves in the Max-tree (this operator will be called kmax), being each node corresponds to a connected component. The results are compared with the optimal thresholding and the Hmaxima technique.
150
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6064A: Image Processing: Algorithms and Systems V
6064A-18, Session 4 Shape-adaptive DCT for denoising and image reconstruction A. Foi, K. Dabov, V. Katkovnik, K. O. Egiazarian, Tampere Univ. of Technology (Finland) The shape-adaptive DCT (SA-DCT) can be computed on a support of arbitrary shape, but retains a computational complexity comparable to that of the usual separable block DCT. Despite the near-optimal decorrelation and energy compaction properties, application of the SA-DCT has been rather limited, targeted nearly exclusively to video compression. It has been recently proposed by the authors to employ the SA-DCT for still image denoising. We use the SA-DCT in conjunction with the directional LPA-ICI technique, which defines the shape of the transform's support in a pointwise adaptive manner. The thresholded or modified SA-DCT coefficients are used to reconstruct a local estimate of the signal within the adaptive-shape support. Since supports corresponding to different points are in general overlapping, the local estimates are averaged together using adaptive weights that depend on the region's statistics. In this paper we further develop this novel approach and extend it to more general restoration problems, with particular emphasis on image deconvolution. Simulation experiments show a state-of-the-art quality of the final estimate, both in terms of objective criteria and visual appearance. Thanks to the adaptive support, reconstructed edges are clean, and no unpleasant ringing artifacts are introduced by the fitted transform. 6064A-19, Session 4 Anisotropic filtering with nonlinear structure tensors C. A. Castaсo Moraga, J. Ruiz-Alzola, Univ. de Las Palmas de Gran Canaria (Spain) The main goal of anisotropic filtering schemes is to reduce noise at the same time that data structure is neither delocalized nor blurred. This idea, which seems very simple at a first glance, requires a lot of care when trying to put it into practice. Literature shows different approaches based on linear filters, PDEs or more sophisticated methods. Nevertheless, all of them are inspired on a common idea: to penalize smoothing along the direction of maximum signal variation while it is favored in the orthogonal one, which is the one of minimum signal variation. For instance, take the example of anisotropic approaches of linear filters. For these filters either the filter coefficients are estimated taking into account the directions of signal variation or the kernel shape is dynamically adapted to the local image features. In this paper, we propose an image processing algorithm which uses both choices. To dynamically adapt the neighborhood to the image features, we propose the use of nonlinear structure tensor. This tensor was recently proposed in the literature and it is a generalization of the classical local structure tensor, where isotropic gaussian smoothing is substituted by a nonlinear diffusion process. In this way, a superior performance to determine signal orientation is obtained at the same time that the typical problems of gaussian smoothing, such as structure delocalization and blurring effect, are almost negligible. Thus, as structure delocalization is less blurred we will be able to better adapt the neighborhoods than using only the linear structure tensor. Size and shape of the neighborhood are estimated using the information provided by the eigenvectors and eigenvalues of this nonlinear structure tensor, which determine an ellipse associated to the quadratic form defined by the positive semidefinite structure tensor. In presence of boundaries or a predominant direction of signal variation, this ellipse is elongated along the orthogonal direction, that is, the one of minimum signal variation. In this way, samples inside the ellipse can be used to perform the estimation and we avoid to mix information from different areas of the image. On the other hand, under no predominant signal variation
the ellipse becomes rounder, so an isotropic estimation is achieved. This idea is quite general and it can be applied to any kind of linear filters, such as Wiener or constrained LMMSE, among others, setting the basis of an anisotropic filtering framework. Hence, to determine the filter coefficients we use an anisotropic gaussian kernel, also driven by the nonlinear structure tensor which weighs more the samples found along the direction of minimum signal variation. Some encouraging results are also presented in this paper in order to show the performance of this approach from a qualitative point of view. A detailed analysis of the results reveals the superior performance of the proposed algorithm in comparison to other adaptive or anisotropic linear filtering schemes, since problems like noise reduction along boundaries or corner delocalization are reduced. In addition, we have set a good framework to be able to determine if the advantages of the nonlinear structure tensor have a positive influence on the filtered image. To do that, we compare the estimated images using the nonlinear structure tensor and the linear counterpart, showing significative differences which allow us to obtain remarkable conclusions. However, some work is still left on the submitted version of the manuscript, since a quantitative analysis of the results is required. 6064A-20, Session 4 A modified wavelet transformation based method of linear object extraction T. Chen, Univ. of South Carolina Edges can be characterized through the evolution of a wavelet transformation at different scale levels. A two-dimensional wavelet transformation of a given image is proportional to the gradient of a corresponding smoothed image. Each component of a normal twodimensional wavelet transformation is in fact a one-dimensional wavelet transformation in one variable followed by a smoothing processing in the other variable. The modified wavelet transformation of the given image gets rid of the smoothing processing since the magnitude of the wavelet transformation in the center of a linear object may be increased by the big magnitudes of the wavelet transformation along the edges if the smoothing processing is adopted, which makes it hard to isolate the centerline of the linear object. The modified wavelet transformation gives high magnitudes along the edges and low magnitudes in the center of the linear objects in the wavelet-transformed image. In the image showing the magnitude of the wavelet transformation, there are high ridges along the edges of the linear objects and low grey level valleys bounded by the ridges. A suitable threshold can be used to extract the low grey level part of the image, such that the centerlines of the linear objects are included. 6064A-21, Session 4 2D approaches to 3D watermarking: state of the art and perspectives M. P. Mitrea, S. A. Duta, F. J. Preteux, Institut National des Tйlйcommunications (France) When dealing with property right protection, watermarking becomes a useful solution, which can afford at the same time reliable owner identification and successful media pirate tracking down, while keeping the same commercial value for the considered piece of media (image, video, audio, 3D). Watermarking can find a theoretical model within the communication theory framework, namely the noisy channel model. Image, video and audio signals have already taken advantage of such a theoretical model, with outstanding result. However, the 3D object watermarking has somehow been left behind. In order to bridge this gap, some studies tried to derive some virtual images (2D data) from the 3D objects. The present paper starts by establishing the overall performances and limitations of such approaches in terms of
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
151
Conf. 6064A: Image Processing: Algorithms and Systems V
transparency, robustness, and data payload. The reason for which the trade off among these three constraints has not yet been reached is determined and discussed. In order to solve this problem, an original 2D/3D watermarking method is further presented. The mark is embedded by an informed embedding approach. The object representation is a set of NURBS (Non Uniform Rational B Spline) surfaces. Firm results concerning robustness (with respect to both 2D and 3D specific attacks) and transparency are obtained, while increasing data payload (up to 20 bits embedded into a quite simple 3D object). 6064A-22, Session 4 Region-based perceptual grouping: a cooperative approach based on DempsterShafer theory N. Zlatoff, B. Tellez, A. M. Baskurt, Univ. Claude Bernard Lyon 1 (France) As segmentation step does not allow to recover semantic objects, perceptual grouping is often used to overcome segmentation's lacks. This refers to the ability of human visual system to impose structure and regularity over signal-based data. Gestalt psychologists have exhibited some properties which seem to be at work for perceptual grouping and some implementations have been proposed by computer vision. However, few of these works model the use of several properties in order to trigger a grouping, even if it can lead to an increase in robustness. We propose a cooperative approach for perceptual grouping by combining the influence of several Gestalt properties for each hypothesis. We make use of Dempster-Shafer formalism, as it can prevent conflicting hypotheses from jamming the grouping process. 6064A-23, Session 5 Study of muscular deformation based on surface slope estimation M. Carli, M. Goffredo, M. Schmid, A. Neri, Univ. degli Studi di Roma Tre (Italy) In this contribution, a novel technique to estimate the skin elasticity factor by using image processing techniques is presented. In human motion analysis, the evaluation of soft tissue deformations, both for clinical and for training purposes, is a relevant task. In non pathological conditions, muscles change shape and size due to contraction, and produce both skin tissue deformations and body shape alterations. In case of skin or muscular pathologies, the skin deformation is not natural, presenting non elastic areas: modelling body limbs with non rigid structures appears a striking solution to highlight this phenomenon. In this paper reflectance models to estimate soft-tissue surface slope are exploited. Skin and muscular deformations are evaluated by analyzing video sequences: the soft tissue modeling is accomplished by using triangular meshes automatically adapting to the body segment. Starting from previous works of the authors the novel approach uses a non linear operator to locally highlight the spatial differences of luminance due to the contraction gesture. A colour map is then associated to the local expansion or contraction of each triangle. The proposed method has been successfully tested on several videos recorded during isometric contractions of biceps brachial.
6064A-24, Session 5 An automated diagnosis approach based on histopathological images M. C. d'Ornellas, C. C. Danesi, J. A. T. Borges da Costa, Univ. Federal de Santa Maria (Brazil) In traditional cancer diagnosis, pathologists examine biopsies to make diagnostic assessments largely based on cell morphology and tissue distribution. However, this is subjective and often leads to considerable variability. On the other hand, computational diagnostic tools enable objective judgments by making use of quantitative measures. This paper presents a systematic survey of the computational steps in automated cancer diagnosis based on histopathology. These computational steps are: 1.) image preprocessing to determine the focal areas, 2.) feature extraction to quantify the properties of these focal areas, and 3.) classifying the focal areas as malignant or not or identifying their malignancy levels. In Step 1, the focal area determination is usually preceded by noise reduction to improve its success. In the case of cellular-level diagnosis, this step also comprises nucleus/cell segmentation. Step 2 defines appropriate representations of the focal areas that provide distinctive objective measures. In Step 3, automated diagnostic systems that operate on quantitative measures are designed. After the design, this step also estimates the accuracy of the system. In this paper, we detail these computational steps, address their challenges, and discuss the remedies to overcome the challenges, emphasizing the importance of constituting benchmark data sets. Such benchmark data sets allow comparing the different features and system designs and prevent misleading accuracy estimation of the systems. Therefore, this allows determining the subsets of distinguishing features, devise new algorithms, and improve the success of automated cancer diagnosis. Today, cancer constitutes a major health problem. Approximately one out of every two men and one out of every three women get cancer at some point during their lifetime. Furthermore, the risk of getting cancer has been further increasing due to the change in the habits of people in our century such as the increase in tobacco use, deterioration of dietary habits, and lack of activity. Fortunately, the recent advances in medicine have significantly increased the possibility of curing cancer. However, the chance of curing cancer primarily relies on its early diagnosis and the selection of its treatment depends on its malignancy level. Therefore, it is critical for us to detect cancer, distinguish cancerous structures from the benign and healthy ones and identify its malignancy level. Traditionally, pathologists use histopathological images of biopsy samples removed from patients, examine them under a microscope, and make judgments based on their personal experience. While examining such images, a pathologist typically assesses the deviations in the cell structures and/or the change in the distribution of the cells across the tissue under examination. However, this judgment is subjective, and often leads to considerable variability. To circumvent this problem and improve the reliability of cancer diagnosis, it is important to develop computational tools for automated cancer diagnosis that operate on quantitative measures. Such automated cancer diagnosis facilitates objective mathematical judgment complementary to that of a pathologist, providing a second opinion for patients. Over the last two decades, a tremendous amount of research work has been conducted for automated cancer diagnosis. This is partly because automated cancer diagnosis holds great promise for largescale use in the advanced cancer treatment and partly because automated cancer diagnosis is not a straightforward task, with a number of challenges to be overcome. The first challenge is the noise elimination in the task of determining the focal areas in the image. The noise arises from staining the biopsy samples; uneven distribution of stain usually cause problems in processing the stained material. In the case of focusing on the properties of nuclei/
152
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6064A: Image Processing: Algorithms and Systems V
cells in the image, the second challenge is the nucleus/cell segmentation. This is challenging because of the complex nature of the image scenes (e.g., touching and overlapping cells) and the noise (e.g., stain artifacts). The third challenge is the feature selection to represent a cell/tissue in the task of cellular or tissuelevel property quantification. The features should provide distinguishing quantitative measures to automatically diagnose the cancer. The last important challenge is the system evaluation in the task of diagnosis. Due to the limited amount of available data, there might be a considerable amount of bias if the system evaluation is not conducted properly. 6064A-25, Session 5 Variational segmentation of x-ray image with overlapped objects G. Yu, Nuctech Co. Ltd. (China); L. Zhang, J. Zhang, Y. Xing, H. Gao, Tsinghua Univ. (China) Image segmentation is a classical and challenging problem in image processing and computer vision. Most of the segmentation algorithms, however, do not consider overlapped objects. Due to the special characteristics of X-ray imaging, the overlapping of objects is very commonly seen in X-ray images and needs to be carefully dealt with. In this paper, we propose a novel energy functional to solve this problem. The Euler-Lagrange equation is derived and the segmentation is converted to a front propagating problem that can be efficiently solved by level set methods. We noticed that the proposed energy functional has no unique extremum and the solution relies on the initialization. Thus, an initialization method is proposed to get satisfying results. The experiment on real data validated our proposed method. 6064A-26, Session 5 Image segmentation for automated dental identification E. Haj Said, D. E. M. Nassar, H. H. Ammar, West Virgina Univ. Dental features are one of few biometric identifiers that qualify for postmortem identification; therefore, creation of an Automated Dental Identification System (ADIS) with goals and objectives similar to the Automated Fingerprint Identification System (AFIS) has received increased attention. As a part of ADIS, teeth segmentation from dental radiographs films is an essential step in the identification process. In this paper, we introduce a fully automated approach for teeth segmentation with goal to extract at least one tooth from the dental radiograph film. The approach is based on performing series of convolution filtering operations using point spread function (PSF) and grouping the pixels based on their connectivity and geometrical properties in order to label each individual tooth. We evaluate our approach based on theoretical and empirical basis, Testing results of segmenting 500 dental radiographs film show that the optimality and the failure rate of our approach are 30.4 % and 1.141% respectively, where the optimality and failure rate percentages capture instances of extreme performance of the segmentation approach. The results show that our approach exhibits the lowest failure rate and the highest optimality among all full automated approaches proposed in the literature.
6064A-27, Session 6 Deblending of the UV photometry in GALEX deep surveys using optical priors in the visible wavelengths A. Llebaria, Lab. d'Astrophysique de Marseille (France); M. Guillaume, D. Aymeric, Ecole Generaliste d'Ingenieurs de Marseille (France); B. Milliard, S. Arnauts, Lab. d'Astrophysique de Marseille (France) Since 2004 the GALEX mission of NASA, is collecting a large set of astronomical UV imaging data in far (~1500A) and near (~2300A) UV wavelengths. The GALEX images have a typical PSF with an FWHM~5arcsec, higher than the optical PSF which reaches currently sub arcsec quality. This difference in resolution introduces ambiguities, miss match and wrong UV flux attribution to the optical counterparts, especially in GALEX deep fields with long exposure time (\>~10000sec). To overcome this problem, we present a new algorithm which allows to measure the UV photometry, through a bayesian approach working in the Poisson noise regime, and using as priors the position and morphological parameters of the optical sources and the GALEX's PSF. This method maximizes the likelihood and generates a set of non-linear equations which are solved using an EM algorithm. For feasibility reasons, since GALEX covers a large field of view (Diam~1.2 deg) with approximately 30000 objects, a segmentation procedure has been defined to manage the analysis in a tractable form. The optical field is thresholded to define a large set of ROIs which are managed with a line adjacency graph (LAG) algorithm. This allows to generate a set of tractable subfields with no interdependency to be analyzed independently in series or concurrently in parallel. In the present paper, we will describe in detail the method and the performances of the UV flux reconstruction algorithm using the optical priors as well as the methodology to handle the original GALEX image and produce the UV source extractions. 6064A-28, Session 6 Comparative study of logarithmic enhancement algorithms with performance measure E. J. Wharton, Tufts Univ.; S. S. Agaian, The Univ. of Texas at San Antonio; K. A. Panetta, Tufts Univ. Performance measures of image enhancement are traditionally subjective and don't quantify the improvement made by the algorithm. Even today, enhancement performance is judged by human observation for many applications. Traditional enhancement measures suffer from the fact that they rely on linear measures of contrast and entropy. This can give preference to images with a large amount of contrast but little useful information because it only considers the total amount of information possible in the image. This could be improved if the measure was biased towards how the human eye sees the image. These shortcomings led to the investigation of the AMEE measure based on entropy. The AMEE based on entropy uses the Michelson contrast, which is better suited for real images than the Weber contrast, which is used by our EME measure. This is because it assumes a periodic pattern as opposed to a small test target at the center of a large uniform background. In this paper, we demonstrate the logarithmic AMEE measure and show how utilizing logarithmic based addition, subtraction, and multiplication provides better results than previously used measures. Finally, we will present a comprehensive study of several enhancement algorithms from all three domains using our measure.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
153
Conf. 6064A: Image Processing: Algorithms and Systems V
6064A-29, Session 6 MMW video sequence denoising and enhancement in concealed weapons detection applications X. Wei, H. Chen, The Univ. of Texas at Arlington; P. K. Varshney, Syracuse Univ. In this paper, we have developed an adaptive algorithm to improve the quality of MMW video sequence by separating the video into foreground region and background region, and handle them differently. We separate the foreground from background area by using an adaptive Kalman filtering. The background is then denoised by both spatial and temporal algorithms. The foreground is denoised by the block-based motion compensational averaging, and enhanced by wavelet-based multi-scale edge representation. Finally further adaptive contrast enhancement is applied to the reconstructed foreground. The experimental results show that the results of our algorithm can achieve smoother background, more reduced noise, more enhanced foreground and higher contrast of the region of interest. 6064A-30, Session 6 Image denoising with block-matching and 3D filtering K. Dabov, A. Foi, V. Katkovnik, K. O. Egiazarian, Tampere Univ. of Technology (Finland) We present a novel approach to still image denoising based on two widely adopted techniques: sliding-window transform processing and block-matching. The latter is predominantly used in video coding to find blocks which follow the motion of objects between consecutive frames. We process blocks within the image in a sliding-window manner and utilize the block-matching concept by searching for variable number of nearby blocks similar to the currently processed one. The matched blocks are stacked together to form a 3D array and due to the similarity between them, the data in the array exhibits high level of correlation. We exploit this correlation by applying a 3D decorrelating unitary transform and attenuate the noise by shrinkage of the transform coefficients. The subsequent inverse transform yields local estimates of the matched blocks. After processing all blocks, the final estimate is a weighed average of all overlapping local ones. We provide experimental results which show that the proposed method delivers highly competitive performance to the state-of-art denoising techniques both in terms of objective criteria and visual quality. 6064A-31, Session 6 An algorithm for the enhancement of images of large dynamic range F. Hassan, J. E. Carletta, Univ. of Akron This paper introduces an algorithm for the enhancement of images of large dynamic range. The algorithm addresses a fundamental problem in digital imaging: the mismatch between the dynamic range of the images as captured by sensors and the dynamic range of the devices on which the images are displayed. This mismatch results in poor quality imaging when images have regions of different illumination; it is difficult to choose a single dynamic range for displaying these images that shows the information in all regions. The proposed algorithm provides not only dynamic range compression, but also separate brightness and contrast control. The algorithm does dynamic range compression by applying a log transformation to the pixel intensity. The transform makes it easier to separate out the illumination and reflectance components of the intensity. We accomplish the separation by applying the biorthogonal 9/7 discrete wavelet transform, which is the basis of the JPEG2000 image compression standard. Illumination information
is concentrated in the approximation subband, while reflectance is concentrated in the details subbands. The algorithm controls brightness by scaling the approximation subband, and contrast by scaling the details subbands. After application of the algorithm, the image can be either reconstructed, or compressed using JPEG2000. 6064A-32, Session 6 Nonlinear image enhancement to improve face detection in complex lighting environment L. Tao, M. Seow, V. K. Asari, Old Dominion Univ. A robust and efficient image enhancement technique has been developed to improve the visual quality of digital images that exhibit dark shadows due to the limited dynamic ranges of imaging and display devices which are incapable of handling high dynamic range scenes. The proposed technique processes images in two separate steps: dynamic range compression and local contrast enhancement. Dynamic range compression is a neighborhood dependent intensity transformation implemented by a nonlinear transfer function defined as the positive side of a hyperbolic tangent function. The slope of the nonlinear function is tuned based on the statistical characteristics of the input image. A neighborhood dependent local contrast enhancement method is then used to compensate the image contrast degraded from the first step by nonlinearly increasing the local intensity variation, which is transformed using a power function with an exponent based on the ratio of the original intensity and the enhanced luminance. Color rendition of the enhanced images is consistent with those of the input image by a linear color re-mapping, which is based on the assumption of consistent ratio between the original and enhanced V component in the HSV color space. Experimental results on the proposed image enhancement technique demonstrates strong capability to improve the performance of convolutional face finder compared to histogram equalization and multiscale retinex with color restoration without compromising the false alarm rate. 6064A-33, Poster Session Noise image enhancement on Hцlder function spaces M. Lee, National Taiwan Ocean Univ. (Taiwan) In this project, we introduce a novel method for restoring noisy images that requires no noisy model. The proposed method is inspired by a wavelet-based switching smoothness description of Hцlder function spaces. We start with a one-dimension signal, expand the signal on a wavelet basis in Hцlder function spaces, and demonstrate how to locally switch the smoothness of a signal. Later, combining the wavelet localization property with adjusting the images Hцlder smoothness, a tool is derived for image enhancement that improves the image quality. We also discuss simulation comparisons with the conventional deconvolution method, which emphasizes the simplicity of the proposed method. Despite the simplicity of our method, significant improvement is reported in the experiment results in terms of image fidelity measure and visual effect. 6064A-34, Poster Session Ad hoc segmentation pipeline for microarray image analysis S. Battiato, G. Di Blasi, G. M. Farinella, G. Gallo, G. C. Guarnera, Univ. di Catania (Italy) DNA microarray is a fundamental biotechnology for genes expression profiling and biomedical studies. Image analysis has
154
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6064A: Image Processing: Algorithms and Systems V
found application in microarray technology because it is able to extrapolate new and not trivial knowledge that is partially hidden in the images. In microarray experiments, two 16-bit TIFF images are obtained by using microarray scanners. The two images are processed to extrapolate data and quality measures by the following steps: gridding, segmentation, intensity extraction and quality measures. Segmentation is a crucial step: it has a potentially large impact on subsequent analysis (e.g. clustering). In this paper we describe MISP (Microarray Image Segmentation Pipeline), a new segmentation pipeline for Microarray Image Analysis. MISP uses a recent segmentation algorithm based on statistical analysis. The Spot masks produced by MISP are used to derive spots information and quality measures. A software prototype system has been developed; it includes visualization, segmentation, information and quality measure extraction. Experiments show the effectiveness of the proposed pipeline both in terms of visual accuracy and measured quality values. Comparisons with existing solutions (e.g. Scanalyze software) confirm the improvement with respect to previously published solutions. 6064A-35, Poster Session Color transient improvement via range detection G. S. Shin, M. G. Kang, Yonsei Univ. (South Korea) In broadcast system, the image information is transmitted in the form of luminance and color difference signals. The color difference signals usually undergo blurs by the several reasons and result in smooth transition. It is important for the CTI algorithm not to produce color mismatch in the smooth transition as well as to make sharp transition. In this paper, the new CTI algorithm which only needs to determine the transition range is proposed. Since the corrected signal does not rely on the high-frequency values, it does not reveal over- and undershoot near edges. To prevent the color mismatch, transition range is found on only one color difference channel. Experimental results show that our algorithm corrects blurred color edges well and is robust to the input images. 6064A-36, Poster Session For fast classification and segmentation of high-resolution images of multiple and complicated colonies W. Wang, Chongqing Univ. of Posts and Telecommunications (China) When colony counting by using an automated image analysis system, it reduces whole cost by approximately 75%. The previous systems, by using an ordinary CCD camera (e.g. b/w camera with resolution up to 768x576), cannot detect the small and tiny colonies (say, the size is less than 0.4 mm). When using the high resolution digital camera, e.g. the image resolution is 2560x1920x24bits (up to 4000x4000x24bits), the information volume is at least 21 times than that from a b/w CCD camera. If we still use the previous process structure or sequence, its low speed cannot meet the automation requirements. This paper presents a methodology for high resolution image classification and segmentation. The size and information volume of the images, taken by a high resolution digital camera, will be tens to hundreds times as the ones taken by an ordinary CCD camera. In order to speed up the image segmentation process of the large images, we classify the images first by using a low resolution image, then, segment them by a fast segmentation algorithm. The algorithm is studied mainly based on multi-resolution technique and the fusion of edge detection result and similarity segmentation result. By use this methodology, the whole image segmentation process time is reduced by tens' times than traditional segmentation methods. And the accuracy of the image segmentation is not decreased.
6064A-37, Poster Session Spatially adaptive multiresolution multispectral image fusion based on Bayesian approach J. H. Park, M. G. Kang, Yonsei Univ. (South Korea) In this paper, we propose two image fusion algorithms based on Bayesian framework for resolution enhancement of remotely sensed multi-spectral images. For both of the algorithms, we acquire a linear MMSE estimator which depends on the first two moments of the joint probability density function (PDF) of the two data. The estimation is performed in a local region which is considered to be stationary. The critical part of the fusion which dominates the performance of algorithms is the estimation process of the first two moments which are used as parameters for fusion. Both of the algorithms consider the local non-stationary nature of images by employing spatially adaptive windows. In the first algorithm, we simplify the process by assuming that the pixels are uncorrelated to each other. To consider the non-linear and spatially non-stationary characteristics of images, we use a spatially adaptive weight function in the parameter estimation. The second algorithm considers the correlation between the pixels in the neighborhood. Through these procedures, the proposed algorithms prevent spectral distortion as much as possible and sharpen the multi-spectral images simultaneously. Experimental results visually illustrate the benefit of the proposed methods when compared to traditional image fusion methods. 6064A-38, Poster Session Segmentation of microspheres in ultrahigh density microsphere-based assays A. Mathur, D. M. Kelso, Northwestern Univ. We have developed a method to identify and localize luminescent microspheres in dense images of microsphere-based assays. Application of this algorithm to the images of densely packed microspheres would aid in increasing the number of assays per unit target sample volume by several orders of magnitude. We immobilize or sediment microspheres on microscope slides and read luminescence from these randomly arrayed microspheres with a digital imaging microscope equipped with a cooled CCD camera. Our segmentation algorithm, which is based on marker-controlled watershed transformation, is then implemented to segment the microsphere clusters in the luminescent images acquired at different wavelengths. This segmentation algorithm is fully automated and require no manual intervention or training sets for optimizing the parameters and is much more accurate than previously proposed algorithms. Using this algorithm, we have accurately segmented more than 97% of the microspheres in dense images. 6064A-39, Poster Session Discrete Gould transform and its applications M. F. Aburdene, H. M. Le, Bucknell Univ. We present a new discrete transform, the Gould transform (DGT) and its applications. The transform has many interesting mathematical properties and potential applications. For example, the forward and inverse transform matrices are both lower triangular, with constant diagonals and sub-diagonals and both can be factored into the product of binary matrices. We applied the transform to edge detection and cryptography. The forward transform can be used to detect edges in digital images. If G is the forward transform matrix and y is the image, then the two dimensional DGT, GyGT can be used directly to detect edges. Ways to improve the technique is to use the "combination of forward and
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
155
Conf. 6064A: Image Processing: Algorithms and Systems V
backward difference", GT(Gy) to better identify the edges. For images that tend to have vertical and horizontal edges, we can further improve the technique by shifting rows (or columns), and then use the technique to detect edges, essentially applying the transform in the diagonal directions. The DGT might have some potential applications in the field of cryptography. 6064A-40, Poster Session The application of image filters combined with the nonlinear regression analysis on optical autofocusing M. Lee, W. Hsu, National Kaohsiung Normal Univ. (Taiwan); T. Lin, National Taiwan Univ. (Taiwan) This paper presents an optical auto-focusing system that is implemented by integrating a real-time auto-focusing algorithm, an image capturing and processing module and a stepper motor. Several image filters are tested and compared through the system for their effects on suppressing noise to accelerate the autofocusing procedure. Besides, a nonlinear regression method is applied in the data analysis for the system to quickly move the stepper motor to the focus. The concise and effective algorithm can be applied on digital cameras for auto-focusing with noise reduction. 6064A-41, Poster Session Crease enhancement and the segmentation of topographic images H. C. Morris, San Josй State Univ. A topographic image can be viewed as a field of ridges and valleys, the gray level representing the altitude; a low gray level is regarded as corresponding to a low altitude, and higher gray levels to higher altitudes. In this paper we present techniques, based on nonlinear diffusion methods, that facilitate the de-noising and segmentation such topographic images. The methods we develop are particularly useful for the analysis of images that display a complex crack structure such as those of planetary surfaces or crystals. There are many works devoted to the extraction of ridge-valley structure. Gaussian de-noising is not appropriate as it destroys the highly oriented textures that we seek to analyze. An alternative is to use anisotropic diffusion filters. However, an undesirable effect of diffusion filters is that ridge-valley junctions are destroyed and nonlinear structures are deformed. In this paper we implement a creaseenhancement diffusion based on the image structure tensor. Our method makes possible to close interrupted linear structures and enhances reliable linear structures. This paper provides a basic review of the mathematics behind crease-enhancing diffusion. It is also shown that the method can also implement a form of grayscale morphology. 6064A-43, Poster Session A multiscale approach to contour detection and texture suppression G. Papari, Rijksuniv. Groningen (Netherlands); P. Campisi, Univ. degli Studi Roma Tre (Italy); N. Petkov, Rijksuniv. Groningen (Netherlands); A. Neri, Univ. degli Studi Roma Tre (Italy) In this paper we propose a multiscale biologically motivated technique for contour detection and texture suppression. Standard edge detectors reacts to all the local luminance changes, irrespective whether they are due to the contours of the objects represented in the scene, rather than to natural texture like grass, foliage, water, etc. Moreover, edges due to texture are often stronger than edges due to true contours. This implies that further process is needed to discriminate true contours from texture edges. In this contribution we exploit the fact that, in a multiresolution analysis, at
coarser scales, only the edges due to object contours are relevant and that texture disappears. This is used in combination with surrounding inhibition, a biologically motivated technique for texture suppression, in order to build a contour detector which is insensitive to texture. Experimental results have also pointed out that our approach is robust to additive noise. 6064A-44, Poster Session A heuristic approach for the extraction of region and boundary of mammalian cells in bio-electric images L. Tao, V. K. Asari, Old Dominion Univ. A robust segmentation and tracking technique for the extraction of region and boundary of mammalian cells in bio-electric images is presented in this paper. The proposed algorithm consists of three steps. The first step is an image enhancement process composed of low-pass filtering and local contrast enhancement. The low-pass filter employs a circular disk operator to suppress the noise. A linear `high-boosting' based algorithm is proposed to enhance the luminance of cell objects with respect to the background. The `degree factor' * introduced for adjusting the contrast degree is the exponential base of the `inverse mean' of the difference of the original and the low-passed image. The second step employs recursive global adaptive thresholding method based on the statistical information of the contrast enhanced image to separate cells from the background. Due to the efficient image enhancement produced in the previous step, global adaptive thresholding is sufficient to provide satisfactory image thresholding results with all cell objects successfully segmented and minimum connections created among adjacent cells. The final step in the segmentation process is composed of boundary tracking and morphological measurement for cell detection. A new robust boundary tracking algorithm using only one searching window with considering boundary tracking route history is proposed to provide rapid and accurate object boundary tracking results. This boundary tracking strategy determines the location of the next boundary pixel by examining the location of the previous boundary pixel and neighboring pixels of the current pixel. This one-window tracking strategy can be summarized and simplified by considering only three cases which will cover all the possibilities without discriminating leftside or right-side searching. This boundary tracking method proves robust performance for any size and complex shaped objects. The tracked boundaries are automatically labeled individually corresponding to the sequence of tracking while they are being searched. Morphological measurements of object perimeter P and area S are obtained after labeling. Mammalian cell decision is implemented based on the areas and the values of form factors of the objects. For the non-cell objects whose areas fall into the range of other cells, the value of object form factor will be considered for comparing to a standard circle. 6064A-45, Poster Session Lip segmentation and tracking for facial palsy M. Park, J. Seo, K. S. Park, Seoul National Univ. (South Korea) We developed the asymmetry analyzing system for facial palsy patient's rehabilitation progress study. Using PC standard imaging device, captured 640*480 RGB image is converted into HSV space. A Lip-shape mask is extracted by thresholding. By taking 5 regions consisted in one region on lip and four regions on face skin, reasonable thresholds are determined by Fuzzy C-Means clustering. The extreme points on the lip shape mask are extracted to get the seeds for tracking. Segmented seed points are tracking by Iterative Lucas-Kanade tracking method in pyramids at 30 fps and recording simultaneously.
156
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6064A: Image Processing: Algorithms and Systems V
To reduce the disk writing load on computer, we use asynchronous mode file writing, which is going to transfer to and review by clinician. Tracking shows quite reliable results, but sometimes the tracked points are following along the lip line because of the similar contrasts. Therefore, the first strategy to improve the reliability of tracking is using the high contrast points, such as left and right maximal point of lip shape. The second is clustering some points near the maximal points and eliminating outlying tracking points. The third is rechecking the lip shape using lip segmentation when the operator confirms that subject's maximal lip moving. Left and right tracking points are compared in forms of trajectory plot. 6064A-46, Poster Session A multithreshold algorithm based on intensity distribution D. Chen, M. Sarhadi, Brunel Univ. (United Kingdom) In current automatic threshold algorithms, the global feature statistic information is used to obtain the threshold to segment or to convert images into binary formats in vision inspection. The main shortcoming of all above algorithms is that they are often designed to process ideal images in laboratory conditions and are not suitable for real world images resulting from industrial environments with uneven illumination distributions. A new automatic threshold algorithm based on multi-thresholds is presented against uneven illumination distribution for the fibre material surface analysis. The proposed multi-threshold algorithm first automatically classifies an image into several geometrical regions that depends on the intensity distribution of the image. Secondly in each region of the image, Otsu's threshold algorithm is applied to obtain the local threshold value. Then the local threshold value is used to segment the local region into the binary format. Finally each local region with its binary format is consisted of the total binary format of the image. Experimental results obtained for images from fibre material in the presence of uneven illumination distribution, show reduced noise levels and enhanced fibre surface properties. The paper presents resulting images and provides key data. 6064A-47, Poster Session An efficient multiresolution GA approach to dental image alignment D. E. M. Nassar, M. Ogirala, D. A. Adjeroh, H. H. Ammar, West Virginia Univ. Automating the process of postmortem identification of individuals using dental records is receiving increased attention in forensic science. Dental image comparison requires proper alignment of teeth before extracting the features used for matching. In this paper we presented an efficient multi-resolution genetic algorithm (MR-GA) approach to the problem of teeth alignment. We use location and orientation attributes of edge points as alignment features and we to seek an affine mapping that aligns a subject tooth to a reference tooth. To search a solution point in the 6D space corresponding to the affine parameters, we instrument GA search progressively across multi-resolution versions of the query and reference images. We use Hausdorff distance to evaluate the fitness of the chromosomes associated with the possible solution points. To efficiently search the parameters space we use some space bounds and thus exclude potentially unreasonable alignment parameters. We tested our MR-GA approach using 52 teeth-pair images and found that our algorithm converges to reasonable solutions in more than 85% of the test cases. The main source of error in the remainder of the cases is the excessive misalignment between the reference and the query images, which is hard to correct within reasonable search time.
6064A-48, Poster Session Phase unwrapping by means of finite differences L. I. Olivos-Pйrez, E. de la Rosa Miranda, L. R. Berriel-Valdos, R. Ramos-Lуpez, Instituto Nacional de Astrofнsica, Уptica y Electrуnica (Mexico) Many problems in metrology and optical tomography have to recover information from phase. In most of the cases, phase, that is associate to a physical magnitude, is continuous and generally, varies smoothly and is wrapping. Therefore, we can say that the problem in these cases is reduced to find a continuous phase. Considering this, many solutions to this kind of problems have been proposed, from the use of local planes to the implementation of most robust algorithms. However, these methods are also very slow. That is why the unwrapping problem is an open subject research in optics. We propose a phase unwrapping method based on finite differences that is fast and robust. In addition, it is easy to program.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
157
Conf. 6064B: Applications of Neural Networks and Machine Learning in Image Processing X
Tuesday-Wednesday 17-18 January 2006 Part of Proceedings of SPIE Vol. 6064 Image Processing: Algorithms and Systems, Neural Networks and Machine Learning
6064B-65, Poster Session Key-text spotting in documentary videos using Adaboost M. Lalonde, L. Gagnon, CRIM (Canada) Important progress regarding text detection in video has been achieved over the last ten years, driven, in particular, by the development of tools for automatic content-based video indexing. This paper presents a method for spotting text in videos based on a cascade of classifiers trained with Adaboost. A video is first segmented into shots, which are then reduced to a set of keyframes. Each keyframe is then analyzed for its text content (subtitles, captions, text that is `naturally' present in the video such as commercial or street signs, etc.). Text spotting is performed by scanning the image with a variable-size window (to account for scale) inside which simple features (mean/variance of grayscale values and x/y derivatives) are extracted in various subareas. Training builds classifiers using the most discriminant spatial combinations of features for text detection. The text-spotting module output is a decision map of the size of the input keyframe showing regions of interest that may contain text suitable for recognition by an OCR system. Test results on documentary films provided by the National Film Board of Canada are given. 6064B-67, Poster Session Research on classifying performance of SVMs with basic kernel in HCCR L. Sun, YanTai Univ. (China) It still is a difficult task for handwritten chinese character recognition (HCCR) to put into practical use. An efficient classifier occupies very important position for increasing offline HCCR rate. SVMs offer a theoretically well-founded approach to automated learning of pattern classifiers for mining labeled data sets. As we know, the performance of SVM largely depends on the kernel function. In this paper, we investigated the performance of SVMs with various common kernels such as linear kernel, polynomial kernel, sigmoid kernel, Gaussian kernel, multi-quadratic kernel, etc. in HCCR. We found that when the feature dimension of character is 64 or lower than 64, Gaussian and multi-quadratic kernel is better than others. When the feature dimension is around 256, the behavior of linear kernel, polynomial kernel, Gaussian and multi-quadratic kernel is pretty much the same thing. 6064B-68, Poster Session Face recognition based on HMM in compressed domain H. Wang, G. Feng, Sun Yat-sen Univ. (China) In this paper we present an approach for face recognition based on Hidden Markov Model (HMM) on compressed domain. We model each individual as an HMM which consists of several face images. A set of DCT coefficients as observation vectors which are obtained from original images by a slipped window are clustered by K-means method using to be the feature of face images. These classified features are applied to train HMMs, equaling to get the parameters of HMMs. Based on the constructed HMMs, face recognition from unknown images are carried out by adjusting experiment parameters (size of window and set of DCT coefficients). ORL face database of 40 individuals with 10 images per individual and Yale face database of 15 individuals with 11 images per individual are used to evaluate the performance of the proposed HMM-based method. For both
ORL and Yale face databases, experiments for different number of images per individual are extracted for training, and the rest ones are for recognition. Different number of images for training corresponds to different experiment parameters. Compared to the other methods relevant to HMM methods reported so far on the two face databases, results of this approach give a better recognition rate, with reduced computational complexity. 6064B-69, Poster Session Application of ANN and DT on medium resolution ASTER image to model gully network in Southern Italy A. Ghaffari, P. M. Mather, G. Priestnall, M. L. Clarke, The Univ. of Nottingham (United Kingdom) This paper describes an approach of utilising ASTER imagery, surface modelling and land cover information to detect gully erosion networks with maximum obtainable accuracy. Comparing and contrasting two different methods, ANN, and Decision Trees (DT) are applied at final stages of research. However, a Grey Level CoOccurrence Matrix texture analysis method (GLCM) has been applied for ASTER bands as one of the input layers. GLCM outputs have just used to combine with geomorphological input layers such as flow accumulation, slope angle and aspect, which have been derived from an ASTER-based Digital Elevation Model (DEM). The ASTER-based DEM with 15-meter resolution has been prepared from L1A. ANN and DT have been used to classify of input layers for 5 sample areas, to the gullies area and none- gully area. Subsequently, The final result shows that DT could classify the image with the highest accuracy (85% overall accuracy) in compare with the ANN. WEKA has used for the analysis as a Machine Learning software. It is possible to improve the accuracy of gully detection using other input layers and classification methods, which could apply during next stages. 6064B-49, Session 7 Nonlinear shape prior from Kernel space for geometric active contours S. Dambreville, Y. Rathi, A. R. Tannenbaum, Georgia Institute of Technology The Geometric Active Contour (GAC) framework, which utilizes image information, has proven to be quite valuable for performing segmentation. However, the use of image information alone often leads to poor segmentation results in the presence of noise, clutter or occlusion. The introduction of shapes priors in the contour evolution proved to be an effective way to circumvent this issue. Recently, an algorithm was proposed, in which linear PCA (principal component analysis) was performed on training sets of data and the shape statistics thus obtained were used in the segmentation process. This approach was shown to convincingly capture small variations in the shape of an object. However, linear PCA assumes that the distribution underlying the variation in shapes is Gaussian. This assumption can be oversimplifying when shapes undergo complex variations. In the present work, we propose to use Kernel PCA to introduce shape prior in the GAC framework. Several experiments were performed using different training-sets of shapes. Starting with any initial contour, we show that the contour evolves to adopt a shape that is faithful to the elements of the training set. The proposed method compares very advantageously to linear PCA.
158
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6064B: Applications of Neural Networks and Machine Learning in Image Processing X
6064B-50, Session 7 Kernel subspace matched target detectors H. Kwon, N. M. Nasrabadi, Army Research Lab. In this paper, we compare several detection algorithms that are based on spectral matched (subspace) filters. Nonlinear kernel versions of these spectral matched (subspace) detectors are also discussed and their performance is compared with the linear versions. These kernel-based detectors exploit the nonlinear correlations between the spectral bands that are ignored by the conventional detectors. Several well-known matched detectors, such as matched subspace detector, orthogonal subspace detector, spectral matched filter and adaptive subspace detector (adaptive cosine estimator) are extended to their corresponding kernel versions by using the idea of kernel-based learning theory. In kernel-based detection algorithms the data is implicitly mapped into a high dimensional kernel feature space by a nonlinear mapping which is associated with a kernel function. The detection algorithm is then derived in the feature space which is kernelized in terms of the kernel functions in order to avoid explicit computation in the high dimensional feature space. Experimental results based on simulated toy-examples and real hyperspectral imagery show that the kernel versions of these detectors outperform the conventional linear detectors. 6064B-51, Session 7 Statistical shape analysis using kernel PCA Y. Rathi, S. Dambreville, A. R. Tannenbaum, Georgia Institute of Technology Mercer kernels are widely used for a wide range of image and signal processing tasks like de-noising, clustering, discriminant analysis etc. These algorithms construct their solutions in terms of the expansions in a high-dimensional feature space F. However, many applications like kernel PCA can be used more effectively if a preimage of the projection in the feature space is available. In this paper, we propose a novel method to reconstruct a unique approximate pre-image of a feature vector and apply it for statistical shape analysis. The proposed reconstruction algorithm can find the pre-image for any invertible kernel. We provide some experimental results to demonstrate the advantages of kernel PCA over linear PCA for shape learning, which include, but are not limited to, ability to learn and distinguish multiple geometries of shapes and robustness to occlusions. We also show the different modes of variation of the learned shapes in the feature space. 6064B-52, Session 8 Segmentation and enhancement of digital copies using a new fuzzy clustering method M. N. Ahmed, B. E. Cooper, Lexmark International, Inc. We introduce a new system to segment and label document images into text, halftoned images, and background using a modified fuzzy c-means (FCM) algorithm. Previously, we introduced a penalty function to the FCM algorithm for the neighborhood term, based on the distance between the neighboring feature vectors and the current cluster center. Inspired by Markov Random Field (MRF) image modeling, the new algorithm is formulated by modifying the objective function of the standard FCM algorithm to allow the labeling of a pixel to be influenced by the labels in its immediate neighborhood. The new cost function conforms with MRF neighborhood modeling through the use of cliques. The objective function is minimized when the center pixel has a high membership value in a certain class at the same time that its related cliques possess low membership values in the other classes.
Each pixel is assigned a feature vector, extracted from edge information and gray level distribution. The feature pattern is then assigned to a specific region using the modified fuzzy c-means approach. In the process of minimizing the new objective function, the neighborhood effect acts as a regularizer and biases the solution towards piecewise-homogeneous labelings. Such a regularization is useful in segmenting scans corrupted by scanner noise. 6064B-54, Session 9 2D/3D facial feature extraction B. Sankur, L. Akarun, H. Cinar, A. Ali Salah, Bogaziзi Univ. (Turkey) With the availability of better sensors, 3D information has become an important source for face recognition, yet face registration and normalization both depend on good localization of fiducial points. In this paper, we compare different methods for landmarking nearfrontal faces automatically. Following a coarse-to-fine approach, three novel methods have been introduced to analyze facial features. The first method uses a mixture of factor analyzers to learn Gabor filter outputs. The second method employs a combination of principal component analysis (PCA) and independent component analysis (ICA) features to analyze a Gabor feature set. The last method uses a subset of DCT coefficients for template-based matching. For the latter two methods we use SVM classifiers with polynomial kernel functions. We contrast 2D texture and 3D depth information, and find that the more successful localization methods use 2D gray level images. The 3D information is still beneficial to the 2D system in eliminating the background, thus reducing the time complexity to one third, and in eliminating false alarms. 6064B-56, Session 10 Noniterative neural network learning of an N-dimension curve representing the dynamic history of a time varying pattern C. J. Hu, Southern Illinois Univ. When an image pattern is varying continuously in time, the arrow tip of an N-Dim analog vector representing the image will trace an NDim curve in the N-space. Like the 2-D and 3-D curves, this N-D curve can also be approximated by some broken N-D straight lines joining the extreme points or "maximum curvature points" on the curve. Each extreme point on the N-D curve represents an extreme "static" pattern of the "dynamic" time-varying pattern. Because of the unique convex learning property possessed by the one-layered, noniterative neural network (OLNN), if one uses an OLNN to learn just the extreme patterns alone, any unlearned test pattern falling close to the ND curve can also be recognized immediately by the OLNN. Therefore it is a very efficient learning-recognition system that will not only learn and recognize the static patterns of the object classes, it will also learn and recognize the dynamic pattern variations of each class. This paper will report the general theoretical analysis and design approaches of this dynamic pattern recognition system. 6064B-57, Session 10 Manifold of color perception: color constancy using a nonlinear line attractor V. K. Asari, M. Seow, Old Dominion Univ. In this paper, we propose the concept of manifold of color perception based on an observation that the perceived color in a set of similar color images defines a manifold in the high dimensional space. Such a manifold representation can be learned from a few images of similar color characteristics. This learned manifold can then be used as a basis for color correction of the images having
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
159
Conf. 6064B: Applications of Neural Networks and Machine Learning in Image Processing X
different color perception to the previously learned color. To learn the manifold for color perception, we propose a novel learning algorithm based on a recurrent neural network. Unlike the conventional recurrent neural network model in which the memory is stored in an attractive fixed point at discrete locations in the state space, the dynamics of the proposed learning algorithm represents memory as a line of attraction. The region of convergence to the line of attraction is defined by the statistical characteristics of the training data. We demonstrate experimentally how we can use the proposed manifold to color-balance the common lighting variations in the environment. 6064B-58, Session 10 A novel neural net application for image compression H. S. Soliman, M. Omari, New Mexico Institute of Mining and Technology Adaptive vector quantization (AVQ) is one of the most important modern techniques used in the domain of lossy image compression because of its low training/operating complexity and high image quality and compression ratio. It clusters similar subimage vectors into the same class with a center of mass representative (centroid) that is stored into an indexed codebook. In the image compression process, each input subimage will be encoded into its corresponding centroid's index of size "log2 [codebook's size]". Hence, the larger the codebook (i.e, richer) the better the image quality, yet the lower the compression ratio. Fortunately, doubling the size of the codebook requires only a single bit addition to every index, due to the logarithmic mapping (log2) of centroids encoding. In order to compromise between the critical factors of compression ratio and quality, we introduce two mechanisms: local and universal codebook. In the former, each image will have its own codebook (space overhead for better quality), with the latter having one large codebook for similar domain images (space saving with lower quality). The construction of the universal codebook is a complex issue (it takes days to fully train few images). Moreover, the inability to automatically detect the image group might result in misclassification and consequently, degrade the image quality. Because of the aforementioned problems of the universal codebook, we introduced the local codebook approach. Each input image is used to train its own private codebook that is able to capture the entire image details. Yet, it is counted against the compression ratio since the codebook is concatenated to the image indices file to form the final compressed image. However, the reconstructed image will be of much better quality compared to that of the universal codebook approach. In addition, the local codebook size is considerably smaller than the image itself, which compensates for the inclusion of the codebook inside the compressed file. Based on the AVQ theory, we designed our new Direct Classification (DC) neural net engine for image compression/decompression. It is based on the philosophy of the winner-take-all feature of the Kohonen model, as well as the elasticity/stability feature with only a "single epoch" training cycle of the ART1 model. In the DC model, the training and compression sub-phases are interleaved. The training sub-phase starts by dividing the total subimage vector domain into classes of subimages. The compression sub-phase consists of assigning the corresponding centroid index as the input subimage compression index. In the decompression phase, each index in the compressed file is looked up in the codebook and replaced with its corresponding centroid. Generating a local codebook per each image seems, at the first glance, a very inefficient approach since it serves a single image, and it is integrated in the compressed file, decreasing the compression ratio. Yet, we discovered experimentally that in case of simple non-complicated images (documents), the codebook is small reducing the space overhead, yet the recovered image quality is improved. We are currently investigating the novelty of our approach in the satellite image domain.
6064B-59, Session 10 Toward content-based object recognition with image primitives G. Wang, J. M. Kinser, George Mason Univ. Content-based object recognition is very useful in many applications, such as medical image processing and diagnosis, target identification with satellite remote sensing. For content-based object recognition, the representation of image segments is critical. Although there are already some approaches to represent image shapes, many of them have limitations because of their insensitivity to the deviations of object appearance. In this paper, an approach is proposed by constructing an image primitive database and representing image with a basis set extracted from the database. The cortical modeling is used here to extract the basis set by isolating the inherent shapes within each image from an image database and defines shapes from this basis set. In our approach, image segments are clustered based on similarity in perimeter and size instead of centroid based metrics by employing the fractional power filter, and the clusters are represented in descriptive vectors as signatures and form basis for shape representation. This approach has advantages in sensitivity to the idiosyncratic nature of the distribution of shapes and efficiency. For validation, we selected a large number of images from web sites randomly. The experiments indicate that describing shapes from this basis set is robust to alterations of the shape such as small occlusions, limited skew, and limited range. 6064B-60, Session 10 Translation invariance in a network of oscillatory units A. R. Rao, G. A. Cecchi, C. Peck, J. Kozloski, IBM Corp. One of the important features of the human visual system is that it is able to recognize objects in a scale and translational invariant manner. However, achieving this desirable behavior through biologistically realistic networks is a challenge. It has been recognized (von der Malsburg [1] ) that neurons are oscillatory dynamical units, and that the synchronization of neuronal firing patterns could be used to solve the binding problem. Networks consisting of such oscillatory units have been applied to solve the signal deconvolution or blind source separation problems. However, the use of the same network to achieve properties that the visual sytem exhibits, such as scale and translational invariance have not been fully explored. Some approaches investigated in the literature (Wallis [2]) involve the use of non-oscillatory elements that are arranged in a hierarchy of layers. The objects presented are allowed to move, and the network utilizes a trace learning rule, where a time averaged value of an output value is used to perform Hebbian learning with respect to the input value. This is a modification of the standard Hebbian learning rule, which typically uses instantaneous values of the input and output. In this paper we present a network of oscillatory amplitude-phase units connected in two layers. The types of connections include feedforward, feedback and lateral. . The network consists of amplitude-phase units that can synchronize their dynamics. We have previously shown (paper submitted to NIPS 2005) that such a network can segment the components of each input object that most contribute to its classification. Learning is unsupervised and based on a Hebbian update, and the architecture is very simple. We extend the ability of the network to address the problem of translational invariance. We show that by a specific treatment of the phase values of the output layer, limited translational invariance is achieved. The scheme used in training is as follows. The network is presented with an input, which then moves. During the motion the
160
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6064B: Applications of Neural Networks and Machine Learning in Image Processing X
amplitude and phase of the upper layer units is not reset, but continues with the past value before the introduction of the object in the new position. Only the input layer is changed instantaneously to reflect the moving object. This is a promising result as it uses the same framework of oscillatory units, and introduces motion to achieve translational invariance. There appears to be some biological grounding for this type of computation, as there are microsaccades during which no phase resetting occurs. It is only when a completely new saccade is undertaken, while looking at a different visual field, that phase resetting occurs. References: [1] J. Buhmann and C. Von Der Malsburg (1991) Sensory segmentation by neural oscillators. International Joint Conference on Neural Networks, Part II, pp. 603-607. [2] Guy Wallis, Using spatio-temporal correlations to learn invariant object recognition, Neural Networks, Dec1996, pp 1513-1519. 6064B-61, Session 10 Efficient learning and recognition using segmented analytical data of an edgedetected 2D image C. J. Hu, Southern Illinois Univ. As we published last year in this conference, we have developed a very efficient image pre-processing scheme for use in any image analyzing system or any neural network pattern recognition system. This scheme calls for a compact mathematical analysis on the binary curves obtained from any original image by a particular scheme, e.g., the Canny edge-detection scheme, which are generally all the binary edge points of the original colored picture. Our preprocessing scheme has a unique property that it can analyze these edge points and synthesize them into jointed branches of analytical curves represented by extremely compact analog data file. Then we can use this compact data file to re-construct, in real time, all the edge boundaries of the objects included in the original image. The error in the re-construction is seen to be very small in all 12 experiments we did last year. This paper reports a noniterative neural network system that will accept this compact data file directly. Then it will automatically analyze the image geometry with possible environmental noises eliminated and with live variation ranges of each pattern taken into account. Finally, it will automatically identify the image according to its overall topological, or syntactic, properties which are generally independent of any geometrical shape, geometrical size, different viewing angles, and different viewing heights. 6064B-62, Session 11 Support vector machine as digital image watermark detector P. H. H. Then, Swinburne Univ. of Technology (Malaysia); Y. C. Wang, Univ. Malaysia Sarawak (Malaysia) We perceive the digital watermark detection as classification problem in image processing. We classify watermarked images as positive class whilst unwatermarked images as negative class. Support Vector Machine (SVM) is used as classifier of the watermarked and unwatermarked digital images. We use two watermarking schemes i.e. Cox's spread spectrum and Single Value Decomposition (SVD) to embed watermark into digital images. The payload of the watermark used for both algorithms is consistent at certain number of bits. SVM is trained with both the watermarked and unwatermarked images. Receiver Operating Characteristics (ROC) graphs are plotted to assess the statistical detection behavior
of both the correlation detector and SVM classifier. We found that straight forward application of SVM leads to generalization problem. We suggest remedies to preprocess the training data in order to achieve substantially better performance from SVM classifier than those resulting from the straightforward application of SVM. Both watermarked and unwatermarked images are attacked under Stirmark and are then tested with the correlation detectors and SVM classifier. A comparison of the ROC of the correlation detectors and SVM classifier is performed to assess the accuracy of SVM classifier relative to correlation detectors. We found that SVM classifier has higher robustness to Stirmark attacks. 6064B-63, Session 11 Neural networks approach to high vertical resolution atmospheric temperature profile retrieval from spaceborne high spectral resolution infrared sounder measurements D. Jiang, C. Dong, Hunan Meterological Bureau (China) Temperature profiles with 1km vertical resolution at 100 pressure layers, from surface up to 0.005 hPa, were retrieved on different spectral bands and on different types of terrain in the middle latitude area by using a three-layered feed-forward neural networks with back-propagation algorithm. Results show that temperature profiles with accuracy of less than 1K in 1 km thick tropospheric layers can be achieved by using AIRS data and neural networks method. And the Qinghai-Tibet Plateau has a measurably impact on the retrieval accuracy which is corresponding to the spectral bands used in performing retrievals. A promising approach to the elimination of this effect is to apply additional predictors which are non-satellite observed (e.g. surface altitude). 6064B-64, Session 11 Probabilistic multiresolution human classification H. Ran, Wuhan Univ. of Technology (China); J. Tu, Hubei Univ. of Technology (China) Recently there has been some interest in using infrared cameras for human detection because of the sharply decreasing prices of easy to use infrared cameras. The training data used in our work for developing the probabilistic template consists of 1000 128x48 rectangular images all of which are known to contain humans in different poses and orientation but having the same height. Multiresolution decomposing is performed on the templates as described above. This is done so that the model does not learn the intensity variations among the background pixels and intensity variations among the foreground pixels. Each template at every level is then translated so that the centroid of the non-zero pixels matches the geometrical center of the image. After this normalization step, for each pixel of the template, the probability p(x,y) of it being pedestrian is calculated based on the how frequently it appears as 1 in the training data. Once we have this probability map, the mean and standard deviation of the combinedprobability is calculated for 1000 training samples. Also the mean and standard deviation is calculated for 1000 128x48 windows that do not contain pedestrians. The videos had quite a lot of variations in the scenes, sizes of people, amount of occlusions and clutter in the backgrounds as is clearly evident. Preliminary experiments shows the robustness.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
161
Conf. 6065: Computational Imaging IV Monday-Wednesday 16-18 January 2006 Part of Proceedings of SPIE Vol. 6065 Computational Imaging IV
6065-01, Session 1 Keynote: title needed M. V. de Hoop, Purdue Univ. No abstract available 6065-02, Session 2 Modeling hierarchical structure of images with stochastic grammars W. Wang, T. Wong, I. Pollak, C. A. Bouman, M. P. Harper, Purdue Univ. We construct a novel hierarchical stochastic model for images. The model is based on the concept of stochastic context-free grammars (SCFGs) which have previously been used for modeling onedimensional objects such as sentences. We extend SCFGs to multiple dimensions, to enable modeling of multidimensional random fields. We develop efficient estimation algorithms for this model and illustrate through several image classification and segmentation examples. 6065-03, Session 2 Multiresolution analysis of digital images using the continuous extension of discrete group transforms M. Germain, J. Patera, A. Zaratsyan, Univ. de Montrйal (Canada) A new technique is presented for multiresolution analysis of digital images. In 2D, it has four variants, two of them are applicable on square lattices like the Discrete Cosine Transform (DCT) which is the simpler. The other two variants would be used similarly on triangular lattices. The property of the Continuous Extension of theses Discrete Group Transforms (CEDGT) is used to analyse data for each level of decomposition. The multiresolution step is obtained by increasing the data grid for each level of decomposition, and by using an adapted low filter to reduce some irregularities due to noise effect. Compared to some stationary wavelet transforms, the image analysis gives better results with a multiresolution CEDGT transform. In fact, a wavelet transform is capable of providing a local representation at multiple scales, but some local details disappear due to the using of the low pass filter and the reduction of the spatial resolution for a high level of decomposition. The smooth interpolation, used by the multiresolution CEDGT, gives interesting results for coarse-to-fine segmentation algorithm and others analysis processes. This technique is applied in segmentation of polarimetric SAR images and the characterization of man-made objects. 6065-04, Session 2 Modeling multiscale differential pixel statistics with applications D. Odom, P. Milanfar, Univ. of California/Santa Cruz Assumptions about the statistical properties of pixel differences between neighboring pixels in an image are often used in image processing tasks. It has been asserted that pixel differences are described well with a Generalized Gaussian distribution. Assuming a Laplace distribution between neighboring pixels (e.g. the Total Variation framework) has also been shown to yield good results. However, neither assumption has been investigated in a statistical
framework. In this paper, the quality of these assumptions was investigated by means of chi-squared goodness of fit tests. A mixture distribution consisting of a Laplace and a Gaussian distribution is proposed as an alternative description of pixel difference statistics. This mixture distribution is seen to provide the best fit to real image data among the distributions tested. A denoising application is presented that utilizes the mixture distribution in the form of a prior. The observed improvement in denoising performance confirms that the mixture model is statistically a better description of real image data. 6065-05, Session 2 Graph-based 3D object classification S. Baloch, A. H. Krim, North Carolina State Univ. No abstract available 6065-06, Session 2 Compression via optimal basis selection in large tree-structured dictionaries Y. Huang, Purdue Univ. No abstract available 6065-07, Session 3 Compressed sensing in noisy imaging environments J. Haupt, R. Castro, R. D. Nowak, Univ. of Wisconsin/Madison Compressive sampling (CS), or "Compressed Sensing", has recently generated a tremendous amount of excitement in the image processing community. CS involves taking a relatively small number of non-traditional samples in the form of projections of the signal onto random basis elements or random vectors (random projections). Recent results show that such observations can contain most of the salient information in the signal. It follows that if a signal is compressible in some basis, then a very accurate reconstruction can be obtained from these random projections. We have shown that compressible signals can be accurately recovered, at the usual nonparametric rates, from random projections that are contaminated with zero-mean additive noise. In many cases this reconstruction is much more accurate than is possible using an equivalent number of conventional point samples, illustrating the utility of projection sampling. For certain classes of sparse or compressible signals, the subspace in which the signal lies may be completely unknown a priori. CS provides universality in such setups because of its ability to both locate and estimate the relevant signal components. In this work we will motivate the use of CS for imaging, present theory predicting reconstruction error rates, and illustrate the utility and universality of CS in electronic imaging with several examples. 6065-08, Session 3 Stable signal recovery from incomplete and inaccurate observations J. K. Romberg, California Institute of Technology Recently, a series of exciting results have shown that it is possible to reconstruct a sparse signal exactly from a very limited number of linear measurements. If our underlying signal f can be written as a superposition of M elements from a known basis, it is possible to recover f from a projection onto a generic subspace of dimension about M log N.
162
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6065: Computational Imaging IV
We will show that the recovery is robust. That is, if the measurements are perturbed, the reconstruction is still stable. We will discuss the implications of this results for applications such as signal compression and tomographic imaging. 6065-09, Session 3 A compressed sensing camera: new theory and an implementation using digital micromirrors D. Takhar, M. B. Wakin, M. Duarte, D. Baron, K. F. Kelly, R. G. Baraniuk, Rice Univ. Current imaging devices rely on CCD or CMOS technology for the optical sensing element. While the scale and cost of this technology have been continually decreasing, the complexity and power requirements have not scaled similarly. On the other hand, digital micromirror devices have proven to be a commercially viable MEMs technology for the video/projector display market. Inspired by the success of this technology, we propose to combine a microcontrolled mirror with a single optical sensor so that it can additionally acquire images, rather than merely adapt current camera technology to serve as an optical sensor. In this project, we have developed a prototype image/video camera based on this concept and realized it through the use of Compressed Sensing for signal reconstruction. Our design has additional desirable properties including scalable output bit stream, variable image resolutions and video frame rates, and an automatically encrypted bit stream at no extra computational or energy cost. 6065-10, Session 4 A fast algorithm for 3D reconstruction from unoriented projections and cryo-electron microscopy of viruses J. Lee, Y. Zheng, P. C. Doerschuk, Purdue Univ. In a cryo electron microscopy experiment, the data is noisy 2-D projection images of the 3-D electron scattering intensity where the orientation of the projections is not known. In previous work we have developed a solution for this problem based on a maximum likelihood estimator that is computed by an expectation maximization algorithm. In the expectation maximization algorithm the expensive step is the expectation which requires numerical evaluation of 3- or 5-dimensional integrations of a square matrix of dimension equal to the number of Fourier series coefficients used to describe the 3-D reconstruction. By taking advantage of the rotational properties of spherical harmonics, we can reduce the integrations of a matrix to integrations of a scalar. The key properties is that a rotated spherical harmonic can be expressed as a linear combination of the other harmonics of the same order and that the weights in the linear combination factor so that each of the three factors is a function of only one of the Euler angles describing the orientation of the projection. 6065-11, Session 4 Spatially adaptive 3D inverse for optical sectioning D. V. Paliy, V. Katkovnik, K. O. Egiazarian, Tampere Univ. of Technology (Finland) One can see well only the focused areas of a three-dimensional (3D) object, observing it in a microscope or another optical device, while others are seen as blurred. However, these out-of-focus structures are in the field of view and thus obscure the in-focus areas. We propose a novel approach to reconstruction of the 3D object from 2D blurred and noisy observations. The technique is based on an approximate image formation model which takes into account depth
varying nature of the blur described by a matrix of shift-invariant 2D point-spread functions (PSF) of an optical system. The proposed restoration scheme incorporates matrix regularized inverse and matrix regularized Wiener inverse algorithms in combination with a novel spatially adaptive denoising technique. This technique is based on special statistical rules for selection of the adaptive size and shape neighbourhood used for the local polynomial approximation of the 2D image intensity. The simulation shows efficiency of the developed approach. 6065-12, Session 4 On soft clipping of Zernike moments for deblurring and enhancement of optical point spread functions N. Becherer, J. Hesser, Univ. Mannheim (Germany) Blur and noise originating from the physical imaging processes degrade the data. Deblurring however demands for an accurate estimation of the underlying point-spread function (PSF). Zernike polynomials allow a compact representation of PSFs since their low order coefficients represent typical aberrations of optical wave fronts while noise is represented by higher order coefficients. A quantitative description of the distribution of noise (Gaussian, Poisson) over the Zernike moments of various orders is given which is the basis for the new soft clipping approach for denoising of PSFs. Instead of discarding moments beyond a certain order, those Zernike moments that are more sensitive to noise are dampened according to the measured distribution and the present noise model. Further, a new scheme to combine experimental and theoretical PSFs in Zernike space is presented. According to our experimental reconstructions, using the new improved PSF the correlation between reconstructed and original volumes is raised by 15% on average cases and up to 85% in the case of thin fibre structures, compared to reconstructions where a non improved PSF was used. Finally we illustrate our results on real life microscopy volumes of cells where our techniques lead to visually improved results. 6065-13, Session 4 Adaptive sampling for atomic force microscopy with system level motion constraints H. Cheng, G. T. C. Chiu, Purdue Univ. In atomic force microscope experiment, a 3-D image of a substrate is obtained. With the total number of samples remains constant, there is a trade-off between the size of the scanned image and the resolution. For the scanning mechanism, the time needed to image an area depends mainly on the number of samples and the size of the image. It is desirable to improve the imaging speed with limited impact to the effective resolution of the portion of the substrate that is of interested. To improve the imaging speed, there are two options: 1) increase the data process rate or 2) reduce the amount of data. Increasing the data process rate is difficult without hardware modifications. Therefore, reducing the amount of data is a more practical and cost effective approach. One key issue for reducing the amount of data is to maintain acceptable image fidelity. To address this issue, we need to classify the sample area into regions based on importance. For high importance regions, a higher resolution is needed. For regions of less importance, a coarse sample density is employed. In this study, we propose a new adaptive sampling scheme that is leveraged from image compression. By adapting the sampling resolution to the substrate profile, the proposed method can decrease the scanning time by reducing the amount of data while maintaining the desired image fidelity.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
163
Conf. 6065: Computational Imaging IV
6065-14, Session 5 Bayesian image reconstruction from Fourier-domain samples using prior edge information: convergence and parameter sensitivity T. S. Denney, Jr., S. J. Reeves, Auburn Univ. Image reconstruction from Fourier-domain measurements is a specialized problem within the general area of image reconstruction using prior information. The structure of the equations in Fourier imaging is challenging, since the observation equation matrix is nonsparse in the spatial domain but diagonal in the Fourier domain. Recently, the Bayesian image reconstruction with prior edges (BIRPE) algorithm has been proposed for image reconstruction from Fourier-domain samples using edge information automatically extracted from a high-resolution prior image. In the BIRPE algorithm, computing the maximum a posteriori (MAP) estimate of the reconstructed image and edge variables involves highdimensional, non-convex optimization, which can be computationally prohibitive. Consequently, we iteratively update our estimate of the image then update our estimate of the edge variables. In this paper, we propose two techniques for updating the image based on fixed edge variables * one based on iterated conditional modes (ICM) and the other based on Jacobi iteration. ICM is guaranteed to converge, but, depending on the structure of the Fourier-domain samples, can be computationally prohibitive. The Jacobi iteration technique is more computationally efficient but does not always converge. In this paper, we study the convergence properties of the Jacobi iteration technique and its parameter sensitivity. 6065-15, Session 5 Thin digital imaging systems using focal plane coding A. D. Portnoy, J. P. Guo, N. P. Pitsianis, D. J. Brady, Duke Univ.; M. A. Fiddy, Univ. of North Carolina/Charlotte; M. R. Feldman, R. D. Te Kolste, Digital Optics Corp. The Compressive Optical MONTAGE Photography Initiative (COMPI) is an effort under the DARPA MONTAGE program to construct thin digital imaging systems while maintaining image quality metrics. COMP-I uses "focal plane coding" to produce nondegenerate data between subapertures. Subaperture data is integrated to form a single high resolution image. Multiple apertures generate multiple copies of a scene on the detector plane. Placed in the image plane, the focal plane mask applies a unique code to each of these sub-images. Within each sub-image, each pixel is masked so that light from only certain optical pixels reaches the detector. Thus, each sub-image measures a different linear combination of optical pixels. Image reconstruction is achieved by inversion of the transformation performed by the imaging system. Registered detector pixels in each sub-image represent the magnitude of the projection of the same optical information onto different sampling vectors. Without a coding element, the imaging system would be limited by the spatial frequency response of the electronic detector pixel. The small mask features allow the imager to broaden this response and reconstruct higher spatial frequencies than a conventional coarsely sampling focal plane.
6065-16, Session 5 3D reconstructions from spherically averaged Fourier transform magnitude and solution x-ray scattering experiments Y. Hwang, P. C. Doerschuk, Purdue Univ. Measuring the scattering a beam of x-rays off a solution of identical particles gives data that is the spherically-averaged magnitude of the Fourier transform of the electron number density in the particle. Although the 1-D data provides only limited information for a 3-D reconstruction of the particle, this approach is still attractive because it does not require that the particle be crystallized for x-ray crystallography or frozen for cryo electron microscopy. We describe ongoing work using two mathematical models of the particle, a piecewise constant model and an orthonormal expansion model, and a variety of specialized optimization tools to determine the 3-D reconstruction of the particle from a weighted nonlinear least squares problem. 6065-17, Session 5 Computed spectroscopy using segmented apertures R. T. Hoctor, F. W. Wheeler, GE Global Research; E. B. Barrett, Lockheed Martin Corp. A novel technique for optical imaging spectroscopy is introduced. The technique makes use of an adjustable aperture and panchromatic light intensity sensors on the imaging plane. The approach is an image reconstruction technique, through which a collection of pan-chromatic images of a scene, each with its own distinct point spread function (PSF), is used to compute a single hyperspectral image (HSI). We have given the name Computed Spectroscopy (CS) to this type of processing. The paper introduces and analyzes Computed Spectroscopy in the context of segmented-aperture optical imaging systems, in which the PSF is adjusted by modifying the path lengths between the image formation plane and a set of independent optical subapertures. Such systems have recently been proposed for Fourier transform imaging spectroscopy (FTIS) applications, and they have long been considered for incoherent synthetic aperture imaging. Computed Spectroscopy is both a new form of computed imaging and an alternative to FTIS that is applicable to division-of-wavefront interferometers. Additionally, the approach provides interesting theoretical connections between image restoration, incoherent aperture synthesis and interferometric spectroscopy. The paper presents analysis and simulation, and it will discuss the limitations and advantages of the new approach. Future research directions will also be discussed. 6065-18, Session 5 Preconditioned conjugate gradient without linesearch: a comparison with the halfquadratic approach for edge-preserving image restoration C. Labat, J. Idier, Institute of Research in Communications and Cybernetics of Nantes (France) Our contribution deals with image restoration. The adopted approach consists in minimizing a penalized least squares (PLS) criterion. Here, we are interested in the search of efficient algorithms to carry out such a task. The minimization of PLS criteria can be addressed using a halfquadratic approach (HQ). However, the nontrivial inversion of a linear system is needed at each iteration. In practice, it is often proposed
164
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6065: Computational Imaging IV
to approximate this inversion using a truncated preconditioned conjugate gradient (PCG) method. However, we point out that theoretical convergence is not proved for such approximate HQ algorithms, referred here as HQ+PCG. In the proposed contribution, we rely on a different scheme, also based on PCG and HQ ingredients and referred as PCG+HQ1D. General linesearch methods ensuring convergence of PCG type algorithms are difficult to code and to tune. Therefore, we propose to replace the linesearch step by a truncated scalar HQ algorithm. Convergence is established for any finite number of HQ1D sub-iterations. Compared to the HQ+PCG approach, we show that our scheme is preferable on both the theoretical and practical grounds. 6065-19, Session 6 Computational methods for image restoration, image segmentation, and texture modeling G. Chung, T. M. Le, L. H. Lieu, N. Tanushev, L. Vese, Univ. of California/Los Angeles This work is devoted to new computational models for image segmentation, restoration and decomposition. In particular, we partition an image into piecewise-constant regions using energy minimization and curve evolution approaches. Applications to brain imaging and tomography will be presented. Also, we decompose a natural image into a cartoon or geometric component, and an oscillatory or texture component, in a variational approach. New computational methods to model oscillatory patterns will be presented. 6065-20, Session 7 An adaptive model for restoration of optically distorted video frames D. Li, Georgia Institute of Technology; M. J. T. Smith, Purdue Univ.; R. M. Mersereau, Georgia Institute of Technology Atmospheric turbulence is a common problem in astronomy and long distance surveillance applications. It can lead to optical distortions that can significantly degrade the quality of the captured images and video. Quality improvement can be achieved through digital restoration methods that effectively suppress the effects of optical distortion. In this paper, atmospheric optical distortion is modeled as having two components: a dispersive component and a time-varying distortion component. A new restoration algorithm is introduced that compensates for dispersion using a fourth-order statistic and employs a new adaptive warping algorithm to suppress turbulent motion effects. The new algorithm is able to improve quality significantly and is able to handle difficult cases involving panning, zooming, and natural motion. 6065-21, Session 7 Resource-driven content adaptation Y. Lu, D. S. Ebert, E. J. Delp III, Purdue Univ. No abstract available 6065-22, Session 7 Algebraic methods for structure from motion M. Boutin, J. Zhang, D. G. Aliaga, Purdue Univ. Structure from motion (SFM) is the problem of reconstructing the geometry of a scene from a stream of images with tracked features.
In this paper, we consider a projective camera model and assume that the internal parameters of the camera are known. Our goal is to reconstruct the geometry of the scene up to a rigid motion (i.e. Euclidean reconstruction.) It has been shown that estimating the pose of the camera from the images is an ill-conditioned problem, as variations in the camera orientation and camera position cannot be distinguished. Unfortunately, the camera pose parameters are an intrinsic part of current formulations of SFM. This leads to numerical instability in the reconstruction of the scene. Using algebraic methods, we obtain a new formulation of SFM which eliminates this cause of instability. 6065-23, Session 7 A maximum entropy kernel density estimator with applications to function interpolation and texture segmentation N. Balakrishnan, D. Schonfeld, Univ. of Illinois at Chicago In this paper, we develop a new algorithm to estimate an unknown probability density function given a finite data sample using a tree shaped kernel density estimator. The algorithm formulates an integrated squared error based cost function which minimizes the quadratic divergence between the kernel density and the Parzen density estimate. The cost function reduces to a quadratic programming problem which is minimized within the maximum entropy framework. The maximum entropy principle acts as a regularizer which yields a smooth solution. A smooth density estimate enables better generalization to unseen data and offers distinct advantages in high dimensions and cases where there is limited data. We demonstrate applications of the hierarchical kernel density estimator for function interpolation and texture segmentation problems. When applied to function interpolation, the kernel density estimator improves performance considerably in situations where the posterior conditional density of the dependent variable is multimodal. The kernel density estimator allows flexible non parametric modeling of textures which improves performance in texture segmentation algorithms. We demonstrate performance on a text labeling problem which shows performance of the algorithm in high dimensions. The hierarchical nature of the density estimator also enables multiresolution solutions depending on the complexity of the data. The algorithm is fast and has at most quadratic scaling in the number of kernels. 6065-24, Session 7 Multiple watermarking: a vector space projection approach O. Altun, G. Sharma, M. Bocko, Univ. of Rochester We present a new paradigm for the insertion of multiple watermarks in images. Instead of an explicitly defined embedding process, the watermark embedding is achieved implicitly by determining a feasible image meeting multiple desired constraints. The constraints are designed to ensure that the watermarked image signal is visually indistinguishable from the original and produces a positive detection result when subjected to detectors for the individual watermarks even in the presence of signal processing operations, particularly compression. We develop useful mathematical definitions of constraint sets for different visual models, for transform domain compression, and for both spread-spectrum and quantization index modulation (QIM) watermark detection scenarios. Using the constraints with a generalized vector space projections method (VSPM), we determine a watermarked signal. Experimental results demonstrate the flexibility and usefulness of the presented methodology in addressing multiple watermarking scenarios while providing implicit shaping of the watermark power to meet visual requirements.
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
165
Conf. 6065: Computational Imaging IV
6065-25, Session 8 Spherical harmonics for shape-based inverse problems, as applied to electrical impedance tomography S. Babaeizadeh, D. H. Brooks, Northeastern Univ. Electrical Impedance Tomography (EIT) is a badly posed inverse problem. In a 3-D volume it requires too many parameters to be able to obtain stable estimates with good spatial resolution and accuracy. One approach to such problems that has been presented recently in a number of reports, when the relevant constituent parameters can be modeled as isotropic and piecewise continuous or homogeneous, is to use shape-based solutions. In this work, we report on a method, based on a spherical harmonics expansion, that allows us to parameterize the 3-D geometry using approximation assumptions about the objects which constitute the conductivity inhomogeneities in the interior; for instance, we could assume the general shape of piecewise constant inhomogeneities is known but their conductivities and their exact location and shape are not. Using this assumption, we have developed a 3-stage optimization algorithm that allows us to iteratively estimate the location of the inhomogeneous objects, to find their external boundaries, and to estimate their internal conductivities. The performance of the proposed method is illustrated via simulation in a realistic torso model, as well as via experimental data from a tank phantom. 6065-26, Session 8 3D nonlinear multigrid algorithm for direct reconstruction of chromophore concentrations in diffuse optical tomography J. C. Ye, Korea Advanced Institute of Science and Technology (South Korea) It is well-known that the tissue has a relatively low absorption for near infrared (NIR) signals between 700nm and 1000nm, which can penetrate up to several inches. For spectroscopic applications, there has been growing interests in using NIR light to monitor physiologically relevant tissue parameters such as blood oxygen saturation and total hemoglobin concentration level in vivo. The diffuse optical tomography (DOT) even further increases the scope of spectroscopic characterization to an imaging level such that we can reconstruct cross-section images of the physiological parameters based on measurements of the scattered and attenuated near infrared light. There are three different types of implementation for DOT - continuous wave (CW), time domain (TD) and frequency domain (FD) approaches. Among them, the CW implementation is the most simple. However, it is known that the inverse problem associated with CW DOT does not have unique solutions due to the cross-talk between the scattering image and absorption image. Recently, Corlue et al describe a model-based spectral approach for inversion of CW data that reduces the cross-talk. In this method, the absorption and scattering parameters are given by wavelength dependent parameteric model such that \begin{eqnarray}\label{eq:model} \mu_a(\lambda)=\sum_i \epsilon_i(\lambda) c_i &,&\mu_s'(\lambda)=a\lambda^{-b} \end{eqnarray} where $c_i$ is the concentration of the $i$-th chromophore, and $\epsilon_i(\lambda)$ is the extinction coefficients of the $i$-th chromophore at wavelength $\lambda$; $a$ and $b$ are related to the size, the index of refraction, and the concentration of scatterers in the tissue as well as the index of refraction of the surrounding medium. Using the parametric models and optimized selection of the multiple wavelength, they demonstrated that the cross talk problem has been significantly reduced. Such parametric model has been also used for frequency domain DOT problem to provide more accurate image reconstruction from multiple excitation. While such finite parametric
representation of optical parameters can alleviate the ill-posedness of inverse problems, one of the main technical difficulties associated with multiple wavelength and parametric model is significantly increased computational complexity. Since the relationship between the measurement and the unknown optical parameters is a highly nonlinear, reconstruction poses a very challenging nonlinear inverse problem even in conventional DOT algorithm. In parametric imaging model, we should furthermore obtain separate images of $c_i$'s, $a$ and $b$ directly from multiple wavelength excitation, which makes the computational burden significantly increasing. In order to overcome the complexity, we investigate the 3-D multigrid algorithms for direct reconstruction of the chromophore concentration. In addition to its computational advantages, the global nature of the multigrid optimization tends to more robustly achieve the global minimum, resulting in improved reconstruction quality. The method is evaluated with simulated data following individual variation of oxygen saturation, hemoglobin, and scattering parameters. A finite-difference model of the diffusion equation is used, and the algorithm reconstructs images of five parameters: oxyhemoglobin, deoxyhemoglobin, water fraction, scatter amplitude, and scatter power, with no assumptions on the scatter amplitude or power. The results show that the new multigrid technique is computationally much less expensive and provide more improved reconstruction quality. 6065-27, Session 8 Adaptation of fast marching methods to subcellular modeling A. Chikando, J. M. Kinser, George Mason Univ. It has long been known that the movement of proteins and other signaling molecules within the cell is controlled by a precise mechanism. However, computational modeling of this mechanism is an ongoing effort. The presented research employs fast marching methods to implement a computer simulation of this cellular transport mechanism. The simulation is then used to illustrate the subcellular trafficking of signaling molecules such as calcium during Excitation - Contraction coupling, and to illustrate proteins conformational and structural rearrangements that occur during translocation to their functional positions within the cell. The developed model adequately simulates diffusive calcium wave subjected to the elements of the calcium release and uptake mechanism during Excitation - Contraction coupling, and illustrates proteins folding and unfolding methodically in order to navigate through the complex cellular cytoskeleton. The accuracy of the simulation is assessed through a detailed exploration of the nascent biophysical properties of proteins, which play a major role in the subcellular transport mechanism. That is, in nature proteins retain their simplest conformation, which is the conformation requiring the least amount of energy. This conformation is a function of the biophysical constituency of their environment. By assessing how well our simulation adheres to these biophysical properties of the protein, we obtain a metric that can be used to measure the accuracy of the simulation. 6065-28, Session 8 Machine learning of human responses to images M. N. Wernick, Y. Yang, J. G. Brankov, L. Wei, Illinois Institute of Technology; I. M. El-Naqa, Washington Univ. in St. Louis; N. P. Galatsanos, Univ. of Ioannina (Greece) The human user is an often ignored component of the imaging chain. In medical diagnostic tasks, the human observer is usually the decision-maker. In content-based image retrieval, the human user is the ultimate judge of the relevance of images recalled from a database. We argue that data collected from human observers should be used in conjunction with machine-learning algorithms to
166
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6065: Computational Imaging IV
model and optimize performance in tasks that involve humans. In essence, we treat the human observer as a nonlinear system to be identified. Two examples of this idea will be discussed. In the first, a learning machine is trained to predict the accuracy of human observers in a lesion detection task. In the second, a learning machine is trained to predict human users' perception of the similarity of two images for purposes of image retrieval from a database. In both examples, it is shown that a nonlinear learning machine can accurately identify the nonlinear human system that maps images into numerical values, such as detection performance or image similarity. 6065-29, Session 9 Image reconstruction algorithms for a novel PET system with a half-ring insert D. Pal, J. A. O'Sullivan, H. Wu, M. Janecek, Y. C. Tai, Washington Univ. in St. Louis We have shown that a higher resolution image can be obtained by incorporating an insert ring with high-resolution detectors around a field-of-view inside a scanner. Coincidences are recorded between pair of detectors in the insert (Type II), the insert and the scanner (Type IS) and the scanner (Type SS). Activity in the space enclosed by the half-ring insert contribute to II, IS and SS types of coincidences. Activity outside the half-ring insert contributes to only IS and SS types of coincidences. This leads to full sampling of the entire imaging FOV, with additional high-resolution sampling within the half-ring. Individual data sets of the half-ring insert system do not provide full sampling of the FOV. A maximum likelihood based expectation maximization (ML-EM) algorithm can combine the log-likelihood functions of all three data sets to estimate a single activity distribution. The algorithm was tested on data from Monte Carlo simulation as well as experiment. Attenuation correction and regularization were incorporated in the EM iterations to obtain a quantitatively accurate image estimate. 6065-30, Session 9 Improved sampling of parallel projection in cylindrical PET scanners B. Farsaii, SUNY/Univ. at Buffalo For a PET scanner with circular array of detectors, the width of lineof-response (LOR) decreases as the distance between the LOR and the center increases. The decrease of width of the LOR leads to problem of non-uniform and under sampling of projections. The consequence of non non-uniform sampling is the distortion of high frequency reconstructed images or loss of fine detail. Correcting this non-uniform sampling problem is known as arc-correction. The purpose of this study is to create the best estimate of non-uniformly sampled projections from uniformly spaced set of LOR. Four polynomial type interpolating algorithms: Lagrange, iterative Neville, natural cubic spline and clamped cubic spline are used to get the best estimate of projections. A set of simulated projections based on GE Advance, CTI EXAT HR+ and CTI ECAT EXACT/953B commercial scanner is generated. To be within the theoretical requirement of three sample points for each FWHM, we have increased the number of sample points form 1 to 6 interpolations per detector. The simulated projections are divided into two groups: the first group consists of a set of functions with no added noise and the second group has added Gaussian noise. Each group consists of two sets: the first set has 10 functions of pulses such that f11 has one pulse, f12 has two pulses and so on. In the second set f21 has one triangular pulse, f22 has two triangular pulses and so on. For each group interpolated data is compared to the original data. In addition, two projections of a 20cm FDG filled disk is used for comparison with simulated data, where the first is the raw projection with no interpolation and the second one is an interpolated projection. It is
shown that clamped and natual cubic spline accuracy was superior to other three algorithms in every case but Lagrange outperforms other algorithms for the speed of execution. 6065-31, Session 9 A Bayesian approach to tomography of multiply scattered beams Z. H. Levine, National Institute of Standards and Technology Recently, Levine, Kearsley, and Hagedorn proposed a generalization of generalized Gaussian random Markov field (GGMRF) as developed by Bouman and Sauer. The principal components of the Bouman-Sauer formulation are a quadratic approximation to the loglikelihood assuming a Poisson distribution and a Beer's Law interaction and a prior distribution which penalized deviation of the values in a neighborhood as a user-defined power in the interval (12]. The generalization removes the restriction that the transmission function follows Beer's Law, but instead admits any functional form for the transmission-thickness relation, such as those arising in transmission electron microscopy of thick samples. Judging from the examples, limited-angle tomography is more sensitive to the choice of the transmission function than all-angle tomography. 6065-32, Session 9 Progress in multiple-image radiography M. N. Wernick, J. G. Brankov, Y. Yang, G. Khelashvili, Illinois Institute of Technology; D. Chapman, Univ. of Saskatchewan (Canada); I. Mondal, B. Marquet, Illinois Institute of Technology; Z. Zhong, Brookhaven National Lab. Conventional mammography is one of the most widely used diagnostic imaging techniques, but it has serious and well-known shortcomings, which are driving the development of innovative alternatives. Our group has been developing an x-ray imaging approach called multiple-image radiography (MIR), which shows promise as a potential alternative to conventional x-ray imaging (radiography). Like computed tomography (CT), MIR is a computed imaging technique, in which the images are not directly observed, but rather computed algorithmically. Whereas conventional radiography produces just one image, depicting absorption effects, MIR simultaneously produces three images, showing separately the effects of absorption, refraction, and ultra-small-angle x-ray scattering. The latter two effects are caused by refractive-index variations in the object, which yield fine image details not seen in standard radiographs. MIR has the added benefits of dramatically lessening radiation dose, virtually eliminating scatter degradation, and lessening the importance of compressing the breast during imaging. Progress to date on the MIR technique will be reviewed in this talk, which will focus on the basic physics and signal-processing issues involved in this new imaging method. 6065-33, Session 9 A recursive filter for noise reduction in tomographic imaging J. Thibault, GE Medical Systems; C. A. Bouman, Purdue Univ.; J. Hsieh, GE Medical Systems; K. D. Sauer, Univ. of Notre Dame CT screening and pediatric imaging, among other applications, prompt the development of more efficient techniques to diminish radiation dose to the patient. While many methods are proposed to limit or modulate patient exposure to x-ray at scan time, the resulting data is excessively noisy, and generates image artifacts unless properly corrected. Statistical iterative reconstruction (IR) techniques have recently been introduced for reconstruction of lowdose CT data, and rely on the accurate modeling of the distribution
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
167
Conf. 6065: Computational Imaging IV
of noise in the acquired data. After conversion from detector counts to attenuation measurements, noisy data usually deviates from a physically accurate representation, and limits the ability of IR to generate artifact-free images. This paper introduces a recursive filter for IR that conserves the physical properties of the measured data while pre-processing attenuation measurements. A basic framework for inclusion of detector electronic noise into the statistical modeling for IR is also presented. The results are shown to successfully eliminate streaking artifacts in photon-starved situations. 6065-34, Session 9 A branch-less distance driven projection and backprojection algorithm S. Basu, General Electric Co.; B. De Man, GE Global Research We have previously presented a technique called Distance Driven (DD) projection and backprojection that has good computational properties as well as sufficient accuracy to avoid high frequency artifacts in either projection or backprojection. In the original formulation, we used a loop structure to compute the overlap kernel that is central to the behavior of the algorithm. We have observed that on some architectures, the branching nature of the kernel significantly affects performance. In this paper, we present a reformulation of our Distance Driven projector and backprojector that eliminates the branches in the inner loop, and also enables the reuse of hardware accelerated components (e.g., Application Specific Integrated Circuits that implement pixel driven projection and backprojection with linear interpolation), while preserving the artifact-free nature of the original technique. 6065-35, Session 9 Cupping artifacts analysis and correction for a FPD-based cone-beam CT L. Zhang, H. Gao, Z. Chen, S. Li, Y. Xing, Tsinghua Univ. (China) Cupping artifact is one of the most serious problems in a middle-low energy X-ray Flat panel detector (FPD)-based cone beam CT system. Both beam hardening effects and scatter could induce cupping artifacts in reconstructions and degrade image quality. In this paper, a two-step cupping-correction method is proposed to eliminate cupping: 1) scatter removal; 2) beam hardening correction. By experimental measurement using Beam Stop Array (BSA), the Xray scatter distribution of a specific object is estimated in the projection image. After interpolation and subtraction, the primary intensity of the projection image is computed. The scatter distribution can also be obtained using convolution with a low-pass filter as kernel. In this paper, linearization is selected as beam hardening correction method for one-material object. For twomaterial cylindrical objects, a new approach without iteration involved is present. There are three processes in this approach. Firstly, correct raw projections by the mapping function of the outer material. Secondly, reconstruct the cross-section image from the modified projections. Finally, scale the image by a simple weighting function. After scatter removal and beam hardening correction, the cupping artifacts are well removed, and the contrast of the reconstructed image is remarkably improved.
6065-44, Poster Session A block-iterative deterministic annealing algorithm for Bayesian tomographic reconstruction S. Lee, Paichai Univ. (South Korea) Most common maximum a posteriori approaches in emission tomography involve assumptions on the local spatial characteristics of the underlying source. To explicitly model the existence of anatomical boundaries, the line-process model has been often used. The unobservable binary line processes in this case acts to suspend smoothness constraints at sites where they are turned on. Due to the nonconvextiy of its objective function, however, the line-process model requires annealing algorithms for optimization. Deterministic annealing (DA) algorithms are known to provide an efficient means of handling the problems associated with mixed continuous and binary variable objectives. However, they are still computer intensive and require many iterations to converge. In this work, to make the DA algorithm permit reconstruction in a clinically acceptable time, we use a block-iterative (BI) method, which is derived from the ordered subsets principle. The BI-DA algorithm processes the data in blocks within each iteration, thereby accelerating the convergence speed of the standard DA algorithm by a factor proportional to the number of blocks. According to our experimental results, the BI-DA algorithm with 32 blocks yielded acceptable images with only 10 iterations, which corresponded to the reconstruction obtained with 320 iterations of the standard DA algorithm. 6065-46, Poster Session Deinterlacing in spatial and temporal domain I. Kim, C. Lee, Yonsei Univ. (South Korea) A number of deinterlacing algorithms have been proposed, which can be divided into two categories: spatial interpolation and temporal interpolation. Each technique has its own advantages and limitations. In order to take advantages of both approaches, attempts have been made to combine both methods. In particular, spatial interpolation methods along with a temporal deinterlacing method using motion compensation have been showing promising results. In this paper, we investigate the performance of several spatial interpolation methods when they are used with motioncompensated deinterlacing methods. 6065-47, Poster Session Cosine transform generalized to lie groups SU(2)xSU(2), O(5) and SU(2)xSU(2)xSU(2): application to digital image processing M. Germain, J. Patera, Univ. de Montrйal (Canada) We propose to apply three of the multiple variants of the 2 and 3dimensional of the cosine transform. We consider the Lie groups leading to square lattices, namely SU(2)xSU(2) and O(5) in the 2dimensional space, and the cubic lattice SU(2)xSU(2)xSU(2) in the 3dimensional space. We aim at evaluating the benefits of some Discrete Group Transform (DGT) techniques, in particular the Continuous Extension of the Discrete Cosine Transform (CEDCT), and at developing new techniques that refine image quality : this refinement is called the high-resolution process. This highest quality is useful to increase the effectiveness of standard features extraction, fusion and classification algorithms. All algorithms based on the 2 and 3-dimensional DGT have the advantage to give the exact value of the original data at the points of the grid lattice, and interpolate well the data values between the grid
168
electronicimaging.org · Tel: +1 703 642 9090 · [email protected]
Conf. 6065: Computational Imaging IV
points. The quality of the interpolation is comparable with the most efficient data interpolation which are currently used for purposes of image zooming. In our first application, we use DGT techniques to refine fully polarimetric radar images, and to increase the effectiveness of standard features extraction algorithms. In our second application, we apply DGT techniques on medical images extracted from a system and a Magnetic Resonance Imaging (MRI) system. 6065-48, Poster Session A prioritized and adaptive approach to volumetric seeded region growing using texture descriptors N. J. Backman, Whitworth College; B. W. Whitney, Northern Kentucky Univ.; J. D. Furst, D. S. Raicu, DePaul Univ. The performance of segmentation algorithms often depends on numerous parameters such as initial seed and contour placement, threshold selection, and other region-dependent a priori knowledge. While necessary for successful segmentation, appropriate setting of these parameters can be difficult to achieve and requires a user experienced with the algorithm of the application field. In order to overcome these difficulties, we propose a prioritized and adaptive volumetric region growing algorithm which will automatically segment a region of interest while simultaneously developing a stopping criterion. This algorithm utilizes volumetric texture extraction to establish the homogeneity criterion whereas the analysis of the aggregating voxel similarities will, over time, define region boundaries. Using our proposed approach on a volume, derived from Computed Tomography (CT) images of the abdomen, we segmented three organs of interest (liver, kidney and spleen). We find that this algorithm is capable of providing excellent volumetric segmentations while also demanding significantly less user intervention than other techniques as it requires only one interaction from the user, namely the selection of a single seed voxel. 6065-49, Poster Session A fast MAP-based superresolution algorithm for general motion M. Tanaka, M. Okutomi, Tokyo Insititute of Technology (Japan) Super-resolution is a technique to reconstruct a high resolution image (HRI) by combining multiple low-resolution images (LRIs). MAP-based super-resolution is one of famous algorithms. A huge calculation amount is an open issue of MAP-based super-resolution. We propose a fast MAP-based super-resolution. The MAP-based super-resolution is formulated as an optimization problem with respect to the HRI. The direct solution of the optimization problem is not feasible, since the number of unknown parameters is typically greater than several thousands. Therefore, the MAP-based superresolution solves that optimization problem by an iterative method. For the proposed algorithm, basic five images related to the cost function are defined. The cost function is reformulated to a combination of simple image operations among those five images. Then, the proposed algorithm optimizes the cost function with respect to the HRI in the frequency domain, whereas existing MAP algorithms optimize with respect to the HRI in the spatial domain. Calculation amount comparison verifies that the proposed algorithm has much smaller calculation cost than a typical algorithm. Experiments using real images captured by a hand-held camera are also demonstrated. They show that the proposed algorithm greatly hastens the super-resolution process, reconstructing an identical HRI to the typical algorithm.
6065-50, Poster Session Image deblurring by the combined use of a superresolution technique and inverse filtering Y. Yamada, K. Nakamae, Osaka Univ. (Japan); H. Fujioka, Fukui Univ. of Technology (Japan) The secondary electron images by scanning electron microscopy (SEM) are considered to be blurred by the electron beam profile. The electron beam profile is approximated by a Gaussian profile. This implies that the transfer function of SEM is necessarily bandlimited. As one of supperresolution techniques that seek to recover information beyond the limit, the extrapolation method by error energy reduction is well known. The method is based on a principle of successive energy reduction. However, when the transfer function is the Gaussian function, it is difficult to extrapolate the spectrum of an object beyond the limit by using the extrapolation method. In this study, we try to deblur the image by the combined use of an extrapolation method and inverse filtering. The procedure is as follows. At first, the electron beam profile "h" is estimated from the secondary electron profile for step edge included in the observed image "g". Then the image "g" is transformed into the image "G" in the frequency domain by the Fourier transform and is low-pass filtered by truncating its Fourier transform to zero outside the interval (-e, e). The bandlimited image "G_lim" is obtained. For the estimated electron beam profile "h", the same process is carried out. That is, "h" is transformed into "H" in the frequency domain and is low-pass filtered. The bandlimited blurring function "H_lim" is obtained. Next the bandlimited original image "F_lim" is computed by using inverse filtering or by dividing "G_lim" by "H_lim". It is well known that when "H" has zero or very small values, then the noise related ratio "N/H" could easily dominate the estimate "G/H". N is the Fourier transform of the additive noise term. By limiting the analysis to frequencies near the origin, we reduce the probability of encountering zero values. Lastly, by applying the error energy reduction extrapolation method to "F_lim", we can obtain the original, deblurring image. We applied our proposed method to simulated blurred images. The quality of our results is superior to that of images restored with Wiener filtering technique. 6065-36, Session 10 Estimation of color filter array data from JPEG images for improved demosaicing W. Feng, S. J. Reeves, Auburn Univ. On-camera demosaicing algorithms are necessarily simple and therefore do not yield the best possible images. However, offcamera demosaicing algorithms face the additional challenge that the data has been compressed and therefore corrupted by quantization noise. We propose a method to estimate the original color filter array (CFA) data from JPEG-compressed images so that more sophisticated (and better) demosaicing schemes can be applied to get higher-quality images. The JPEG image formation process, including simple demosaicing, color space transformation, chrominance channel decimation and DCT, is modeled as a series of matrix operations followed by quantization on the CFA data, which is estimate