Maximal granularity structure and generalized multi-view discriminant analysis for person re-identification

作者：

Highlights：

• This paper proposed a new maximal granularity structure descriptor which extracts local salient features.

• To make astable representation against viewpoint changes, this paper presented a novel strategyof crossing coding.

• Besides, considering multi-view information, this paper provided a new GMDA-RC method for different views.

摘要

•This paper proposed a new maximal granularity structure descriptor which extracts local salient features.•To make astable representation against viewpoint changes, this paper presented a novel strategyof crossing coding.•Besides, considering multi-view information, this paper provided a new GMDA-RC method for different views.

论文关键词：Person re-identification,Maximal granularity structure descriptor,Generalized multi-view discriminant analysis,Representation consistency

论文评审过程：Received 10 July 2017, Revised 13 January 2018, Accepted 28 January 2018, Available online 2 February 2018, Version of Record 13 February 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.01.033