Camera-aware representation learning for person re-identification
文献类型:期刊论文
作者 | Wu, Jinlin4,5,6; Yang, Yuxin3; Lei, Zhen4,5,6; Yang, Yang4,5,6; Chen, Shukai1; Li, Stan Z.2 |
刊名 | NEUROCOMPUTING |
出版日期 | 2023-01-21 |
卷号 | 518页码:155-164 |
ISSN号 | 0925-2312 |
关键词 | Camera-imbalanced data distribution Sub -center hard mining Camera -balanced memory bank Multiple -center Softmax |
DOI | 10.1016/j.neucom.2022.11.009 |
通讯作者 | Lei, Zhen() |
英文摘要 | Person re-identification (ReID) aims to associate the same person across non-overlapping cameras. However, most of existing works neglect the issue of camera-imbalanced data distribution. Consequently, pedestrian representation learning gives preference to the head cameras, which have com-paratively more training data, and disregards the tail cameras, which have relatively less training data. In this paper, we propose a novel framework for camera-aware representation learning to overcome this issue, named CARL. On the proxy level representation learning, CARL presents a multiple-center soft -max loss to correct the head camera bias and presents a hard sub-center mining strategy to improve the discrimination of tail camera samples. On the pair-wise level representation learning, CARL builds a camera-balanced memory bank (CBM) to re-balance the sample pair distribution and proposes a camera-paired loss for pair-wise metric learning. Extensive experiments and ablation studies on MSMT17, the current largest ReID dataset with massive camera-imbalanced data distribution, demon-strate that our CARL is superior to previous metric learning based ReID methods and achieves state-of-the-art performance. (c) 2022 Elsevier B.V. All rights reserved. |
资助项目 | National Key Research & Development Program ; Chinese National Natural Science Foundation ; [2020YFC2003901] ; [62276254] ; [62206280] ; [62176256] ; [61976229] ; [62106264] |
WOS研究方向 | Computer Science |
语种 | 英语 |
出版者 | ELSEVIER |
WOS记录号 | WOS:000904670800014 |
资助机构 | National Key Research & Development Program ; Chinese National Natural Science Foundation |
源URL | [http://ir.ia.ac.cn/handle/173211/51159] |
专题 | 多模态人工智能系统全国重点实验室 |
通讯作者 | Lei, Zhen |
作者单位 | 1.ZKTeco Co Ltd, Dongguan, Peoples R China 2.Westlake Univ, Sch Engn, Hangzhou, Peoples R China 3.Sichuan Univ, Chengdu, Peoples R China 4.Univ Chinese Acad Sci, Beijing, Peoples R China 5.Chinese Acad Sci, Inst Automat, NLPR, Beijing, Peoples R China 6.Chinese Acad Sci, Inst Automat, CBSR, Beijing, Peoples R China |
推荐引用方式 GB/T 7714 | Wu, Jinlin,Yang, Yuxin,Lei, Zhen,et al. Camera-aware representation learning for person re-identification[J]. NEUROCOMPUTING,2023,518:155-164. |
APA | Wu, Jinlin,Yang, Yuxin,Lei, Zhen,Yang, Yang,Chen, Shukai,&Li, Stan Z..(2023).Camera-aware representation learning for person re-identification.NEUROCOMPUTING,518,155-164. |
MLA | Wu, Jinlin,et al."Camera-aware representation learning for person re-identification".NEUROCOMPUTING 518(2023):155-164. |
入库方式: OAI收割
来源:自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。