Will scene information help realistic action recognition?

The question of scene information whether can help realistic action recognition has been investigated in this paper. The salience region of each frame in video was acquired by using Itti-Koch algorithm. The information outside the salience region represented scene information. Two action recognition methods were tested on the YouTube action dataset. One method got rid of partial scene information, while the other contained scene information. The obtained impressive results showed that scene information can help realistic action recognition.

[1]  Liu Juan,et al.  Recognizing Realistic Human Actions Using Accumulative Edge Image , 2012 .

[2]  Xu Guang-You,et al.  Viewpoint Independent Action Recognition , 2008 .

[3]  Fei-Yue Huang.  Viewpoint Independent Action Recognition: Viewpoint Independent Action Recognition , 2008 .

[4]  Christof Koch,et al.  A Model of Saliency-Based Visual Attention for Rapid Scene Analysis , 2009 .

[5]  Wei Liang,et al.  Discriminative human action recognition in the learned hierarchical manifold space , 2010, Image Vis. Comput..

[6]  Jiebo Luo,et al.  Recognizing realistic actions from videos “in the wild” , 2009, 2009 IEEE Conference on Computer Vision and Pattern Recognition.

[7]  Andrew Zisserman,et al.  Representing shape with a spatial pyramid kernel , 2007, CIVR '07.

[8]  Christof Koch,et al.  Control of Selective Visual Attention: Modeling the Where Pathway , 1995, NIPS.

[9]  Ronald Poppe,et al.  A survey on vision-based human action recognition , 2010, Image Vis. Comput..

[10]  Cordelia Schmid,et al.  Human Detection Using Oriented Histograms of Flow and Appearance , 2006, ECCV.

[11]  Hao Aimin,et al.  View-Invariant Action Recognition Based on Action Graphs , 2009 .

[12]  S. Sitharama Iyengar,et al.  Histogram-based morphological edge detector , 1994, IEEE Trans. Geosci. Remote. Sens..

[13]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).