Action Recognition from Still Images Based on Deep VLAD Spatial Pyramids

作者：

Highlights：

• A novel method of spatial pyramid VLAD encoding using patches of CNN features is proposed for action recognition from still images.

• Adding a spatial pyramid to VLAD encoding significantly boosts the system performance even VLAD encoding on its own shows improved results compared with CNN feature.

• The method has been validated on four widely used datasets with competitive results to demonstrate this scheme's applicability on action recognition and attribute classification.

摘要

Highlights•A novel method of spatial pyramid VLAD encoding using patches of CNN features is proposed for action recognition from still images.•Adding a spatial pyramid to VLAD encoding significantly boosts the system performance even VLAD encoding on its own shows improved results compared with CNN feature.•The method has been validated on four widely used datasets with competitive results to demonstrate this scheme's applicability on action recognition and attribute classification.

论文关键词：Actions,Convolutional Neural Networks,VLAD encoding,Spatial pyramids

论文评审过程：Received 17 October 2016, Revised 16 March 2017, Accepted 16 March 2017, Available online 18 March 2017, Version of Record 21 March 2017.

论文官网地址：https://doi.org/10.1016/j.image.2017.03.010