Vehicle re-identification based on unsupervised local area detection and view discrimination

作者：

Highlights：

• We propose a novel unsupervised discriminative area detector to localize on local areas of vehicle images.

• We design a multi-branch CNN network to extract the global and the discriminative local features.

• We propose a novel view-discrimination based reranking algorithm to which optimizes recognition results.

摘要

•We propose a novel unsupervised discriminative area detector to localize on local areas of vehicle images.•We design a multi-branch CNN network to extract the global and the discriminative local features.•We propose a novel view-discrimination based reranking algorithm to which optimizes recognition results.

论文关键词：Vehicle re-identification,Unsupervised,Discriminatory local area,View discrimination,Cross-view

论文评审过程：Received 16 May 2020, Revised 27 July 2020, Accepted 23 August 2020, Available online 19 September 2020, Version of Record 24 October 2020.

论文官网地址：https://doi.org/10.1016/j.imavis.2020.104008