Laplacian group sparse modeling of human actions

作者：

Highlights：

• We learn a discriminative representation of human behavior with the structural information among local features.

• Video-level sparsity is promoted for the human behavior representation via Laplacian group sparse coding.

• We get a closed-form solution to reach the objective of our model, which makes the method very efficient.

摘要

•We learn a discriminative representation of human behavior with the structural information among local features.•Video-level sparsity is promoted for the human behavior representation via Laplacian group sparse coding.•We get a closed-form solution to reach the objective of our model, which makes the method very efficient.

论文关键词：Action recognition,High-level representation,Laplacian group sparse coding,Structural information

论文评审过程：Received 14 April 2013, Revised 7 February 2014, Accepted 12 February 2014, Available online 20 February 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.02.007