| Presentation | 2017-10-13 Robust Human Pose Estimation from Distorted Images Daisuke Miki, Shinya Abe, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Predicting a human pose from an image is a technique used for detection of abnormalities. However, existing motion capture systems need complicated imaging devices such as stereo camera or infrared camera. It is also difficult to recognize an individual’s pose from a short distance using the devices in use today. In this study, we present a robust human pose estimation method using a distorted fisheye image with a device that can capture images from a wide-angle as well as at a close-range. We propose a multi-layered convolutional neural network architecture for estimating an individual’s joint positions and deformation parameters to enable robustness to image distortion. We confirmed the ability of wide-angle and close-range detection through the publicly available datasets. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Human pose estimation / Deep learning / Convolutional neural network / Fisheye image |
| Paper # | PRMU2017-93 |
| Date of Issue | 2017-10-05 (PRMU) |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2017/10/12(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Shinichi Sato(NII) |
| Vice Chair | Hironobu Fujiyoshi(Chubu Univ.) / Yoshihisa Ijiri(Omron) |
| Secretary | Hironobu Fujiyoshi(AIST) / Yoshihisa Ijiri(NAIST) |
| Assistant | Masato Ishii(NEC) / Yusuke Sugano(Osaka Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Pattern Recognition and Media Understanding |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Robust Human Pose Estimation from Distorted Images |
| Sub Title (in English) | |
| Keyword(1) | Human pose estimation |
| Keyword(2) | Deep learning |
| Keyword(3) | Convolutional neural network |
| Keyword(4) | Fisheye image |
| 1st Author's Name | Daisuke Miki |
| 1st Author's Affiliation | Tokyo Metropolitan Industrial Technology Research Institute(TIRI) |
| 2nd Author's Name | Shinya Abe |
| 2nd Author's Affiliation | Tokyo Metropolitan Industrial Technology Research Institute(TIRI) |
| Date | 2017-10-13 |
| Paper # | PRMU2017-93 |
| Volume (vol) | vol.117 |
| Number (no) | PRMU-238 |
| Page | pp.pp.169-174(PRMU), |
| #Pages | 6 |
| Date of Issue | 2017-10-05 (PRMU) |