Datum-Adaptive Local Metric Learning for Person Re-identification

作     者：Liu, Kai Zhao, Zhicheng Cai, Anni 

作者机构：Beijing Univ Posts & Telecommun Multimedia Commun & Pattern Recognit Labs Beijing Key Lab Network Syst & Network Culture Beijing 100088 Peoples R China 

出 版 物：《IEEE SIGNAL PROCESSING LETTERS》 (IEEE Signal Process Lett)

年 卷 期：2015年第22卷第9期

页      面：1457-1461页

核心收录：

学科分类：0808[工学-电气工程] 08[工学] 

基　　金：Chinese National Natural Science Foundation [61471049, 61101212, 61372169] National High Technology R&D Program of China (863 Program) [2012AA012505] 

主　　题：Local Coordinate Coding local metric learning person re-identification 

摘      要：Person re-identification (PRID) is a challenging problem in multi-camera surveillance systems. In this paper, we propose a novel Datum-Adaptive Local Metric learning method for PRID, which learns individual local feature projection for each image sample according to the current data distribution and projects all samples into a common discriminative space for similarity measure. We adopt an approximate strategy based on Local Coordinate Coding to learn local projections. Anchor points are first generated by clustering and the local projection of each sample is then approximated by the linear combination of a set of projection bases, which are associated with the anchor points. Experimental results demonstrate that the proposed approach obtains superior performance compared with state-of-the-art methods on public benchmarks.