Action recognition via spatio-temporal local features: A comprehensive study

作者：

Highlights：

• A timely review of action recognition based on spatio-temporal local features.

• Transfer techniques in image domain to video domain for action recognition.

• Evaluation of representation methods for action recognition.

摘要

•A timely review of action recognition based on spatio-temporal local features.•Transfer techniques in image domain to video domain for action recognition.•Evaluation of representation methods for action recognition.

论文关键词：Action recognition,Spatio-temporal local features,Feature coding,Bag-of-words,Sparse coding,Fisher kernel,VLAD,NBNN,Match kernels,Performance evaluation

论文评审过程：Received 5 September 2013, Revised 24 October 2015, Accepted 19 February 2016, Available online 20 March 2016, Version of Record 4 April 2016.

论文官网地址：https://doi.org/10.1016/j.imavis.2016.02.006