From Groups to Co-Traveler Sets: Pair Matching Based Person Re-Identification Framework

Min Cao, Chen Chen, Xiyuan Hu, Silong Peng; Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 2573-2582

Abstract


In video surveillance, group refers to a set of people with similar velocity and close proximity. Group members can provide visual clues for person re-identification. In this paper, we discuss the essentials of group-based person re-identification and relax the group definition towards a concept of "co-traveler set", keeping constraints on velocity differences while loosening the distance constraint. Accordingly we propose a pair matching scheme to measure the distance between co-traveler sets, which tackles the problems caused by dynamic change of group across camera views. The final individual matching score is weighted by the obtained distance measurements between co-traveler sets. A proof of concept shows the rationality of introducing the concept of co-traveler relation into person reid. Experiments were conducted on four different datasets. Our co-traveler set based framework shows promising improvement compared with the group-based methods and the individual-based methods.

Related Material


[pdf]
[bibtex]
@InProceedings{Cao_2017_ICCV,
author = {Cao, Min and Chen, Chen and Hu, Xiyuan and Peng, Silong},
title = {From Groups to Co-Traveler Sets: Pair Matching Based Person Re-Identification Framework},
booktitle = {Proceedings of the IEEE International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2017}
}