Evaluating spatiotemporal interest point features for depth-based action recognition

作者：

Highlights：

• A comprehensive evaluation of STIP based features on depth-based action recognition

• Two schemes to refine STIP features for a deeper understanding of their behaviors

• A fusion approach is developed which outperforms many state-of-the-art methods.

摘要

•A comprehensive evaluation of STIP based features on depth-based action recognition•Two schemes to refine STIP features for a deeper understanding of their behaviors•A fusion approach is developed which outperforms many state-of-the-art methods.

论文关键词：Action recognition,Spatiotemporal interest point (STIP),Detectors,Descriptors,STIP features,RGB-D sensor,Evaluation,STIP feature refinement,Feature fusion

论文评审过程：Received 11 August 2013, Revised 6 January 2014, Accepted 2 April 2014, Available online 16 April 2014.

论文官网地址：https://doi.org/10.1016/j.imavis.2014.04.005