Flow guided mutual attention for person re-identification

作者：

Highlights：

• A Mutual Attention Network is proposed to leverage both Optical FLow and Video Stream input for Video Person ReID.

• Longer sequences help in capturing robust features.

• Longer sequences need careful attention based weighing to dis-regard outliers.

• The Mutual Attention model can be used with different backbones.

摘要

•A Mutual Attention Network is proposed to leverage both Optical FLow and Video Stream input for Video Person ReID.•Longer sequences help in capturing robust features.•Longer sequences need careful attention based weighing to dis-regard outliers.•The Mutual Attention model can be used with different backbones.

论文关键词：Video surveillance,Person re-identification,Optical flow,Metric learning,Attention mechanisms

论文评审过程：Received 27 November 2020, Revised 23 May 2021, Accepted 25 May 2021, Available online 7 July 2021, Version of Record 13 July 2021.

论文官网地址：https://doi.org/10.1016/j.imavis.2021.104246