Infrared-visible cross-modal person re-identification via dual-attention collaborative learning

作者：

Highlights：

• Exchange information among multiple classifiers.

• A collaborative learning strategy with constraints is proposed.

• Mine useful feature through a dual-attention module.

• A new backbone sharing strategy is designed.

摘要

•Exchange information among multiple classifiers.•A collaborative learning strategy with constraints is proposed.•Mine useful feature through a dual-attention module.•A new backbone sharing strategy is designed.

论文关键词：Person re-identification,Cross-modality,Dual attention,Collaborative learning

论文评审过程：Received 9 July 2021, Revised 13 June 2022, Accepted 2 September 2022, Available online 13 September 2022, Version of Record 22 September 2022.

论文官网地址：https://doi.org/10.1016/j.image.2022.116868