Research on action recognition method under mobile phone visual sensor
- 10.2991/amcce-15.2015.211How to use a DOI?
- mobile phone; sensor; action recognition;
During action recognition process under the mobile phone visual sensor, in order to ensure the speed of recognition, generally distributing as sloppy feature points, so as to express composite action feature needs to be recognized. However, if the action is composite feature in small area, traditional action recognition method based on sparse representation, when expressing action, is not able to express feature details within small area in high accuracy and achieve action recognition under mobile phone visual sensor accurately and effectively. To this end, an action recognition method under mobile phone visual sensor based on visual optimization analysis of behavior is proposed. With the maximum similarity clustering method to segment action images obtained by mobile phone vision sensor accurately, to exclude a large number of irrelevant features. By fitting the action clustering function to calculate principal curvature of action edge position, so as to acquire the position and scale of action. By adopting the action gray level co-occurrence matrix to apply adaptive classification for action, quickly complete the accurate identification of the action under mobile phone visual sensor. The simulation results show that, this algorithm has high stability and accuracy.
- © 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 - Wenbin Wang AU - Ketang Chen AU - Liangliang Chen PY - 2015/04 DA - 2015/04 TI - Research on action recognition method under mobile phone visual sensor BT - Proceedings of the 2015 International Conference on Automation, Mechanical Control and Computational Engineering PB - Atlantis Press SN - 1951-6851 UR - https://doi.org/10.2991/amcce-15.2015.211 DO - 10.2991/amcce-15.2015.211 ID - Wang2015/04 ER -