Multiple face tracking and recognition with identity-specific localized metric learning

作者：

Highlights：

• Multiple face recognition and tracking are formulated in a unified integer programming optimization framework.

• The identity-specific localized metrics with adaptive weights are jointly learned online for more robust appearance matching.

• Experimental evaluations demonstrated the superior performance of our method in various challenging recognition scenarios.

摘要

•Multiple face recognition and tracking are formulated in a unified integer programming optimization framework.•The identity-specific localized metrics with adaptive weights are jointly learned online for more robust appearance matching.•Experimental evaluations demonstrated the superior performance of our method in various challenging recognition scenarios.

论文关键词：Multiple face recognition,Multiple face tracking,Integer programming,Localized metric learning

论文评审过程：Received 8 November 2016, Revised 17 July 2017, Accepted 12 September 2017, Available online 23 September 2017, Version of Record 21 November 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.09.022