High performance person re-identification via a boosting ranking ensemble

作者：

Highlights：

• We propose a novel boosting ensemble approach for person Re-id.

• We formulate the ensemble model for person Re-id from a ranking perspective.

• We utilize a rectifier loss to enhance the boosting framework to alleviate overfitting.

• Experimental performance demonstrate the proposed approach achieves superior performance.

摘要

•We propose a novel boosting ensemble approach for person Re-id.•We formulate the ensemble model for person Re-id from a ranking perspective.•We utilize a rectifier loss to enhance the boosting framework to alleviate overfitting.•Experimental performance demonstrate the proposed approach achieves superior performance.

论文关键词：Person re-identification,Boosting,Ensemble

论文评审过程：Received 8 September 2018, Revised 11 April 2019, Accepted 13 May 2019, Available online 25 May 2019, Version of Record 29 May 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.05.022