Person re-identification with multi-level adaptive correspondence models》.pdfVIP

Neurocomputing 168 (2015) 550–559 Contents lists available at ScienceDirect Neurocomputing journal homepage: /locate/neucom Person re-identification with multi-level adaptive correspondence models Shi-Chang Shi a, Chun-Chao Guo a,c, Jian-Huang Lai a,b,n, Shi-Zhe Chen a, Xiao-Jun Hu a a School of Information Science and Technology, Sun Yat-sen University, Guangzhou 510006, P.R. China b Guangdong Key Laboratory of Information Security Technology, Guangzhou 510006, P.R. China c SYSU-CMU Shunde International Joint Research Institute (JRI), Shunde 528300, P.R. China a r t i c l e i n f o a b s t r a c t Article history: In this work, we present a multi-level adaptive correspondence model for person re-identification. Coarse Received 10 December 2014 segmentation and single level representation carry poorly discriminative information for generating a Received in revised form signature of a target, whilst fine segmentation with a fixed matching fashion is hindered severely by 8 April 2015 misalignment of corresponding body parts. We address such a dilemma through a multi-level adaptive Accepted 20 May 2015 correspondence scheme. Our approach encodes a pedestrian based on horizontal stripes in multi-level to Communicated by Rongrong Ji Available online 28 May 2015 capture rich visual cues as well as implicit spatial structure. Then dynamic correspondence of stripes within an