CVPR'99 Tutorials Program June 21-22, 1999 Colorado State University Fort Collins, Colorado, USA Web Site: http://www.cs.colostate.edu/~cvpr99/worktut.html A total of five tutorials will be held in conjunction with the 1999 IEEE Computer Vision and Pattern Recognition Conference. The tutorials program spans the two days before the main CVPR'99 conference. PROGRAM Monday, June 21st, 1999 M1. 3D Photography (9:00am - 5:00 pm) S. Seitz, B. Curless, P. Debevec, M. Levoy, J.-Y. Bouguet M2. Automated Biometrics (8:00am - 12:00pm) N. Ratha, A. Senior, R. Bolle M3. Multiscale Geometric Image Analysis: Scale-Space Theory (1:30pm-5:30pm) B. Romeny Tuesday, June 22nd, 1999 T4. Multiple View Geometry (9:00am - 5:00pm) R. Hartley, A. Zisserman T5. Video computing (8:00am - 12:00pm) M. Shah REGISTRATION Tutorial registration is in the form of One-day or Two-day Tutorial Passports. Passports include admission to any of the tutorials offered on a single day for a One-day Passport, and on both days for a Two-day Passport. To register for tutorials, please use the CVPR'99 registration forms. These can be obtained online from http://www.cs.colostate.edu/~cvpr99/registration.html TUTORIAL ABSTRACTS M1. 3D Photography Organizers: S. Seitz (CMU), B. Curless (U. Washington-Seattle) Additional speakers: P. Debevec (U. California-Berkeley) M. Levoy (Stanford), J.-Y. Bouguet (CalTech) This course provides an introduction to 3D photography: the process of using cameras and light to capture the shape and appearance of real objects. Methods include both passive and active vision techniques ranging from stereo, structure from motion, and photogrammetry to imaging radar, optical triangulation, and interferometry. The course introduces these fundamental methods, provides in-depth analysis of several emerging techniques, and concludes with a field study: capturing 3D photographs of Michelangelo's statues. M2. Automated Biometrics Organizers: N. Ratha (IBM), A. Senior (IBM), R. Bolle (IBM) This tutorial will address many research as well as practical issues in automated biometrics. The underlying pattern recognition and computer vision techniques will be reviewed. The state of the art in fingerprint, face, iris and speaker identification will be presented. Issues in integrating biometrics and representative applications will also be discussed. M3. Multiscale Geometric Image Analysis: Scale-Space Theory Organizer: B. Romeny (Utrecht U.) The tutorial focuses on multiscale image analysis, its relation to human vision, and the differential structure of images in a modern, physics based approach. The course is an introduction and overview of the field of Gaussian scale-space theory, for computer scientists at the graduate level. Goal is primarily to give an intuitive notion of the important physics and mathematics involved, and to give a wide range of instructive applications T4. Multiple View Geometry Organizers: R. Hartley (GE), A. Zisserman (Oxford U.) The key problem to be discussed is the reconstruction of scenes from multiple images. Emphasis will be on an approach using uncalibrated, or partially calibrated cameras and will be directed to a beginning and intermediate-level audience. The aim of this tutorial is twofold: 1. To introduce the key ideas in two, three and N-view camera geometry. This will include the fundamental matrix for two views; the trifocal tensor for 3 views; projective reconstruction from two or more views; auto-calibration from two or more views. 2. To describe algorithms for computing these relations. The tutorial will include numerous laptop demonstrations of these relations including "live" computations from images. T5. Video Computing Organizer: M. Shah (U. Central Florida) During the last three decades, computer vision researchers have worked on methods for analyzing sequences of images (video). This work has mainly focussed on measurement of 2D motion (e.g., optical flow), 3D motion (structure from motion), tracking and scene change detection. Recently, there has been growing interest in recognition of motion from video sequences. The progress has been made in developing several working systems for recognizing hand gestures, human activities and behaviors, lip movements, and facial expressions. Interesting work also has been reported on video mosaics, video phones, video synthesis, video abstraction and retrieval. This tutorial will provide an in depth introduction to various aspects of video. The participants will be able learn important algorithms (with implementation details) in video understanding. The video demonstrations of key systems will be presented, and pointers to the recent papers on this topic will be provided. FOR MORE INFORMATION For detailed outlines of each tutorial, please check the CVPR'99 tutorials web page at http://www.cs.colostate.edu/~cvpr99/worktut.html