Title
Channel-exchanged feature representations for person re-identification
Abstract
•A novel two-stream channel-exchanged network named CENet is proposed for person re-identification.•The CENet can extract color-robust features to alleviate the interference of color in person re-identification.•A data augmentation method is introduced to person re-identification for increasing the training set.•An improved multi-layer structure is applied for further enhancing the Re-ID accuracy.
Year
DOI
Venue
2021
10.1016/j.ins.2021.03.028
Information Sciences
Keywords
DocType
Volume
Person re-identification,Feature representation,Color-robust,Multi-layer learning,Multi-task learning
Journal
562
ISSN
Citations 
PageRank 
0020-0255
0
0.34
References 
Authors
0
5
Name
Order
Citations
PageRank
Jianchen Wang100.34
Liming Yuan243.09
Haixia Xu303.04
Gengsheng Xie400.34
Xian-Bin Wen55516.67