Human and action recognition using adaptive energy images

作者：

Highlights：

• A new representation is proposed for person identification and action recognition using Depth video information and correlation coefficient of motion sequences.

• Using correlation coefficient to generate templates that adaptable according to the completion time of movements was achieved better results.

• The experimental results of proposed method are compared with previous studies successful results are obtained.

• Proposed method, prevents time loss by avoiding empirical methods to find out which highlighted TT is more appropriate for which action.

摘要

•A new representation is proposed for person identification and action recognition using Depth video information and correlation coefficient of motion sequences.•Using correlation coefficient to generate templates that adaptable according to the completion time of movements was achieved better results.•The experimental results of proposed method are compared with previous studies successful results are obtained.•Proposed method, prevents time loss by avoiding empirical methods to find out which highlighted TT is more appropriate for which action.

论文关键词：Motion recognition,Human recognition,Correlation coefficients,Deep learning,Behavioral biometrics

论文评审过程：Received 11 December 2019, Revised 18 February 2022, Accepted 2 March 2022, Available online 3 March 2022, Version of Record 10 March 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108621