Learning Cross-domain Dictionary Pairs for Human Action Recognition
- 10.2991/iwmecs-15.2015.84How to use a DOI?
- Human action recognition, Local motion pattern, Dictionary learning.
This paper present a cross domain dictionary learning way, via the introduction of auxiliary domain, as the extra knowledge, the intra class diversity of the original training set (also known as the target domain) is effectively enhanced. Firstly, use local motion pattern feature as a low-level feature descriptor, and then through a cross domain reconstructive dictionary pair learning, which brings the original target data and the auxiliary domain data into the same feature space to get corresponding sparse codes of each human action categories. Finally, classification and recognition is carried on the human action representation. Using the UCF YouTube data set as the original training set and the HMDB51 data set as the auxiliary data set, the recognition rate of the proposed framework is significantly improved on the UCF YouTube dataset.
- © 2015, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Bingbing Zhang AU - Dongcheng Shi AU - Kang Ni AU - Chao Liang PY - 2015/10 DA - 2015/10 TI - Learning Cross-domain Dictionary Pairs for Human Action Recognition BT - Proceedings of the 2015 2nd International Workshop on Materials Engineering and Computer Sciences PB - Atlantis Press SP - 419 EP - 424 SN - 2352-538X UR - https://doi.org/10.2991/iwmecs-15.2015.84 DO - 10.2991/iwmecs-15.2015.84 ID - Zhang2015/10 ER -