DMCA
Monocular tracking 3d people by gaussian process spatio-temporal variable model
| Venue: | In Proc. IEEE International Conference on Image Processing ICIP 2007 |
| Citations: | 2 - 0 self |
BibTeX
@INPROCEEDINGS{Pang_monoculartracking,
author = {Junbiao Pang and Laiyun Qing and Qingming Huang and Shuqiang Jiang and Wen Gao},
title = {Monocular tracking 3d people by gaussian process spatio-temporal variable model},
booktitle = {In Proc. IEEE International Conference on Image Processing ICIP 2007},
year = {},
pages = {41--44}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Tracking 3D people from monocular video is often poorly constrained. To mitigate this problem, prior knowledge should be exploited. In this paper, the Gaussian process spatio-temporal variable model (GPSTVM), a novel dy-namical system modeling method is proposed for learning human pose and motion priors. The GPSTVM provides a low dimensional embedding of human motion data, with a smooth density function that provides higher probability to the poses and motions close to the training data. The low dimensional latent space is optimized directly to retain the spatio-temporal structure of the high dimensional pose space. After the prior on human pose is learned, the particle filtering can be used tracking articulated human pose; parti-cle filtering propagates over time in the embedding space, avoiding the curse of dimensionality. Experiments demon-strate that our approach tracks 3D people accurately.
Keyphrases
gaussian process spatio-temporal variable model human pose monocular video articulated human pose training data spatio-temporal structure low dimensional latent space particle filtering low dimensional embedding human motion data motion prior parti-cle filtering propagates high dimensional pose space novel dy-namical system smooth density function
