Person Reidentification Via Discrepancy Matrix And Matrix Metric
Keywords
Discrepancy matrix; matrix metric; metric projection; person reidentification (re-id)
Abstract
Person reidentification (re-id), as an important task in video surveillance and forensics applications, has been widely studied. Previous research efforts toward solving the person re-id problem have primarily focused on constructing robust vector description by exploiting appearance's characteristic, or learning discriminative distance metric by labeled vectors. Based on the cognition and identification process of human, we propose a new pattern, which transforms the feature description from characteristic vector to discrepancy matrix. In particular, in order to well identify a person, it converts the distance metric from vector metric to matrix metric, which consists of the intradiscrepancy projection and interdiscrepancy projection parts. We introduce a consistent term and a discriminative term to form the objective function. To solve it efficiently, we utilize a simple gradient-descent method under the alternating optimization process with respect to the two projections. Experimental results on public datasets demonstrate the effectiveness of the proposed pattern as compared with the state-of-the-art approaches.
Publication Date
10-1-2018
Publication Title
IEEE Transactions on Cybernetics
Volume
48
Issue
10
Number of Pages
3006-3020
Document Type
Article
Personal Identifier
scopus
DOI Link
https://doi.org/10.1109/TCYB.2017.2755044
Copyright Status
Unknown
Socpus ID
85031769041 (Scopus)
Source API URL
https://api.elsevier.com/content/abstract/scopus_id/85031769041
STARS Citation
Wang, Zheng; Hu, Ruimin; Chen, Chen; Yu, Yi; and Jiang, Junjun, "Person Reidentification Via Discrepancy Matrix And Matrix Metric" (2018). Scopus Export 2015-2019. 9344.
https://stars.library.ucf.edu/scopus2015/9344