Paper Abstract and Keywords |
Presentation |
2015-03-19 16:10
Articulated pose estimation from ego-centric vision for action recognition Haruka Yonemoto, Kazuhiko Murasaki, Tatsuya Osawa, Kyoko Sudo, Jun Shimamura, Yukinobu Taniguchi (NTT) BioX2014-56 PRMU2014-176 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In this paper, we present a novel method to estimate articulated human pose for ego-centric action
recognition. Many researches of action recognition from third-person vision have shown that articulated human
pose is a feature which directly dscribes human motion, and is invaliant to the view from camera. However, they
assume the whole gure of subjects appears in the image and ,in our best knowlege, no method for estimating ar-
ticulated pose from ego-centric vision is proposed. Our method can extract the pose by integrating hand detection,
camera pose estimation, and time-series ltering with the constraint of body shape. The experimental evaluations
show that joint positions are correctly estimated even though they do not appear in the image, and that the accuracy
of action recognition is improved through motion features based on estimated poses compared with motion features
extracted from only hand position in the image. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
action recognition / articulated pose estimation / ego-centric vision / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 114, no. 521, PRMU2014-176, pp. 105-110, March 2015. |
Paper # |
PRMU2014-176 |
Date of Issue |
2015-03-12 (BioX, PRMU) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
BioX2014-56 PRMU2014-176 |
|