A Deep and Structured Metric Learning Method for Robust Person Re-Identification

doi:10.1016/j.patcog.2019.106995

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 生物识别与安全技术

	A Deep and Structured Metric Learning Method for Robust Person Re-Identification
	Ren, Chuan-Xian 1; Xu, Xiao-Lin 2; Lei, Zhen3,4
发表期刊	PATTERN RECOGNITION
ISSN	0031-3203
	2019-12-01
卷号	96 页码:12
通讯作者	Ren, Chuan-Xian()
摘要	Person re-identification (re-ID) is to match different images of the same pedestrian. It has attracted increasing research interest in pattern recognition and machine learning. Traditionally, person re-ID is formulated as a metric learning problem with binary classification output. However, higher order relationship, such as triplet closeness among the instances, is ignored by such pair-wise based metric learning methods. Thus, the discriminative information hidden in these data is insufficiently explored. This paper proposes a new structured loss function to push the frontier of the person re-ID performance in realistic scenarios. The new loss function introduces two margin parameters. They operate as bounds to remove positive pairs of very small distances and negative pairs of large distances. A trade-off coefficient is assigned to the loss term of negative pairs to alleviate class-imbalance problem. By using a linear function with the margin-based objectives, the gradients w.r.t. weight matrices are no longer dependent on the iterative loss values in a multiplicative manner. This makes the weights update process robust to large iterative loss values. The new loss function is compatible with many deep learning architectures, thus, it induces new deep network with pair-pruning regularization for metric learning. To evaluate the performance of the proposed model, extensive experiments are conducted on benchmark datasets. The results indicate that the new loss together with the ResNet-50 backbone has excellent feature representation ability for person re-ID. (C) 2019 Elsevier Ltd. All rights reserved.
关键词	Metric learning Feature extraction Deep neural networks Imbalance regularization Person re-identification
DOI	10.1016/j.patcog.2019.106995
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[61572536] ; National Natural Science Foundation of China[11631015] ; National Natural Science Foundation of China[U1611265] ; Science and Technology Program of GuangZhou[201804010248] ; National Natural Science Foundation of China[61572536] ; National Natural Science Foundation of China[11631015] ; National Natural Science Foundation of China[U1611265] ; Science and Technology Program of GuangZhou[201804010248]
项目资助者	National Natural Science Foundation of China ; Science and Technology Program of GuangZhou
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS记录号	WOS:000487569700033
出版者	ELSEVIER SCI LTD
引用统计	被引频次：15[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/26627
专题	多模态人工智能系统全国重点实验室_生物识别与安全技术
通讯作者	Ren, Chuan-Xian
作者单位	1.Sun Yat Sen Univ, Sch Math, Intelligent Data Ctr, Guangzhou 510275, Guangdong, Peoples R China 2.Guangdong Univ Finance & Econ, Sch Math & Stat, Guangzhou 510320, Guangdong, Peoples R China 3.Chinese Acad Sci, Inst Automat, Ctr Biometr & Secur Res, Beijing 100190, Peoples R China 4.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Ren, Chuan-Xian,Xu, Xiao-Lin,Lei, Zhen. A Deep and Structured Metric Learning Method for Robust Person Re-Identification[J]. PATTERN RECOGNITION,2019,96:12.
APA	Ren, Chuan-Xian,Xu, Xiao-Lin,&Lei, Zhen.(2019).A Deep and Structured Metric Learning Method for Robust Person Re-Identification.PATTERN RECOGNITION,96,12.
MLA	Ren, Chuan-Xian,et al."A Deep and Structured Metric Learning Method for Robust Person Re-Identification".PATTERN RECOGNITION 96(2019):12.