Multi-type attributes driven multi-camera person re-identification

作者：

Highlights：

• We propose a weakly-supervised attribute learning algorithm that learns attributes from a limited number of labeled data.

• A novel dCNN structure is proposed to predict attributes into multiple types to ensure their incompatibility.

• Deep attributes achieve promising performance and generalization ability in person ReID task.

摘要

•We propose a weakly-supervised attribute learning algorithm that learns attributes from a limited number of labeled data.•A novel dCNN structure is proposed to predict attributes into multiple types to ensure their incompatibility.•Deep attributes achieve promising performance and generalization ability in person ReID task.

论文关键词：Deep attributes,Person re-identification

论文评审过程：Received 25 September 2016, Revised 1 June 2017, Accepted 4 July 2017, Available online 12 September 2017, Version of Record 21 November 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.07.005