Presentation | 2004-12-17 Parametric eigenspace method with sparsely distributed multiple cameras Hidenori Tanaka, Itaru Kitahara, Hideo Saito, Hiroshi Murase, Kiyoshi Kogure, Norihiro Hagita, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In this paper, we propose a learning method of parametric eigenspace using sparsely distributed multiple cameras. Personal identification is one of the most important purposes to install video sensors to the ubiquitous computing environment. However, when the sensor is actually used in the real world, we cannot always capture the desired appearance information, because it is not possible to completely control target objects (e.g. pedestrians). To tackle with this problem, we automatically interpolate the object's appearances which are captured by sparsely distributed multiple cameras through a cylindrical 3D model, and generate an initial eigenspace with them. When a new image is captured, an interpolated image which is the most similar in the eigenspace to the input image is replaced. Repeating this procedure, we can reconstruct the eigenspace with the new data set. Experimental results show that the discernment capability of the initial eigenspace is improved by repeating the updating process. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | approximation 3D model / parametric eigenspace method / learning / view interpolation / multiple videos |
Paper # | PRMU2004-137 |
Date of Issue |
Conference Information | |
Committee | PRMU |
---|---|
Conference Date | 2004/12/10(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Pattern Recognition and Media Understanding (PRMU) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Parametric eigenspace method with sparsely distributed multiple cameras |
Sub Title (in English) | |
Keyword(1) | approximation 3D model |
Keyword(2) | parametric eigenspace method |
Keyword(3) | learning |
Keyword(4) | view interpolation |
Keyword(5) | multiple videos |
1st Author's Name | Hidenori Tanaka |
1st Author's Affiliation | Intelligent Robotics and Communication Laboratories, ATR:Graduate School of Science and Technology, Keio University() |
2nd Author's Name | Itaru Kitahara |
2nd Author's Affiliation | Intelligent Robotics and Communication Laboratories, ATR |
3rd Author's Name | Hideo Saito |
3rd Author's Affiliation | Graduate School of Science and Technology, Keio University |
4th Author's Name | Hiroshi Murase |
4th Author's Affiliation | Graduate School of Information Science, Nagoya University |
5th Author's Name | Kiyoshi Kogure |
5th Author's Affiliation | Intelligent Robotics and Communication Laboratories, ATR |
6th Author's Name | Norihiro Hagita |
6th Author's Affiliation | Intelligent Robotics and Communication Laboratories, ATR |
Date | 2004-12-17 |
Paper # | PRMU2004-137 |
Volume (vol) | vol.104 |
Number (no) | 524 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |