Workshop on Science of Computer Vision

Date: Thursday 19 - Friday 20 September 2002
Venue: Lecture Hall: Graduate School of Natural Science and Technology
Okayama 700-8530, Japan
(Location)
Language: English
Admission: Free (Reception excluded)
Invited talks:: Amnon Shashua (Biography), Hebrew University of Jerusalem, Israel(Abstract)
Fredrik Kahl (Biography), Lund University, Sweden (Abstract)

Program

Thursday 19 September

9:30 - 9:35 Opening remarks: Kenichi Kanatani (Okayama University)
9:40 - 10:45 Invited talk (1): 3D interpretation of dynamic and non-rigid scenes.
Amnon Shashua (Hebrew University of Jerusalem, Israel)
10:50 - 11:30: For geometric inference, what kind of statistical model is necessary?
Kenichi Kanatani (Okayama University)
11:30 - 12:10: Theory of nuisance parameter problems in computer vision and examples of application for linear case.
Takayuki Okatani and Koichiro Deguchi (Tohoku University)
12:10 - 13:30: Lunch
13:40 - 14:20: 3-D Reconstruction from optical flow based on semi-parametric model.
Naoya Ohta (Gunma University)
14:20 - 15:00: Automatic camera model selection for multibody motion segmentation.
Yasuyuki Sugaya and Kenichi Kanatani (Okayama University)
15:00 - 15:40: Robust image matching under a large disparity.
Yasushi Kanazawa (Toyohashi University of Technology) and Kenichi Kanatani (Okayama University)
15:50 - 16:30: Optical flow detection by an extended gradient-based method.
Hidetoshi Miike, Atsushi Osa and Atsushi Nomura (Yamaguchi University)
16:30 - 17:10: Deformation based on perception of the human visual system.
Atsushi Osa, Kazumi Nagata (Yamaguchi University), Sosuke Tsukamoto (National Institute for Longevity Sciences), Takeshi Kinoshita, Makoto Ichikawa and Hidetoshi Miike (Yamaguchi Universtiy)
18:30 - Reception

Friday 20 September

9:40 - 10:45 Invited talk (2): Critical configurations for projective reconstruction.
Fredrik Kahl (Lund University, Sweden)
10:50 - 11:30: Image based visual servoing with partitioned approach.
Koichiro Deguchi (Tohoku University)
11:30 - 12:10: Calibration of multi-camera systems using planar patterns.
Toshio Ueshiba and Fumiaki Tomita (National Institute of Advanced Industrial Science and Technology)
12:10 - 13:30 Lunch
13:40 - 14:20: Projection-based approach to robust face recognition from too few images.
Takeshi Shakunaga, Fumihiko Sakaue and Kazuma Shigenari (Okayama University)
14:20 - 15:00: Constructing illumination image basis from object motion.
Akiko Nakashima, Atsuto Maki and Kazuhiro Fukui (TOSHIBA)
15:10 - 15:50: Conic-based multiple view geometry.
Akihiro Sugimoto (National Institute of Informatics)
15:50 - 16:30: Determining camera intrinsic and extrinsic parameters from multiple images of multiple balls.
Gang Xu (Ritsumeikan University)

Organizer:: Kenichi Kanatani, Professor
Department of Computer Science
Okayama University
Okayama 700-8530, Japan
Phone/Fax: +81(86)251-8173