Exemplar-Based Human Action Pose Correction and Tagging
Abstract
The launch of Xbox Kinect has built a very successful computer vision product and made a big impact to the gaming industry; this sheds lights onto a wide variety of potential applications related to action recognition. The accurate estimation of human poses from the depth image is universally a critical step. However, existing pose estimation systems exhibit failures when faced severe occlusion. In this paper, we propose an exemplar-based method to learn to correct the initially estimated poses. We learn an inhomogeneous systematic bias by leveraging the exemplar information within specific human action domain. Our algorithm is illustrated on both joint-based skeleton correction and tag prediction. In the experiments, significant improvement is observed over the contemporary approaches, including what is delivered by the current Kinect system.
Cite
Text
Shen et al. "Exemplar-Based Human Action Pose Correction and Tagging." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012. doi:10.1109/CVPR.2012.6247875Markdown
[Shen et al. "Exemplar-Based Human Action Pose Correction and Tagging." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012.](https://mlanthology.org/cvpr/2012/shen2012cvpr-exemplar/) doi:10.1109/CVPR.2012.6247875BibTeX
@inproceedings{shen2012cvpr-exemplar,
title = {{Exemplar-Based Human Action Pose Correction and Tagging}},
author = {Shen, Wei and Deng, Ke and Bai, Xiang and Leyvand, Tommer and Guo, Baining and Tu, Zhuowen},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {2012},
pages = {1784-1791},
doi = {10.1109/CVPR.2012.6247875},
url = {https://mlanthology.org/cvpr/2012/shen2012cvpr-exemplar/}
}