A framework for semantic people description in multi-camera surveillance systems

作者：

Highlights：

• No training or camera calibration is needed and achieve real-time processing.

• Extract descriptors from targets with rotations and scale change.

• Descriptors of partially visible people can also be extracted.

• No high resolution cameras are needed.

摘要

•No training or camera calibration is needed and achieve real-time processing.•Extract descriptors from targets with rotations and scale change.•Descriptors of partially visible people can also be extracted.•No high resolution cameras are needed.

论文关键词：People re-identification,Human appearance model,Semantic features,Soft-biometric,Surveillance

论文评审过程：Received 1 July 2014, Revised 25 November 2015, Accepted 28 November 2015, Available online 18 January 2016, Version of Record 1 February 2016.

论文官网地址：https://doi.org/10.1016/j.imavis.2015.11.009