Rank-1 Tensor Approximation for High-Order Association in Multi-target Tracking

doi:10.1007/s11263-018-01147-z

	Rank-1 Tensor Approximation for High-Order Association in Multi-target Tracking
	Shi, Xinchu1 ; Ling, Haibin 4; Pang, Yu 4; Hu, Weiming2,3 ; Chu, Peng 4; Xing, Junliang1
刊名	INTERNATIONAL JOURNAL OF COMPUTER VISION
	2019-08-01
卷号	127 期号:8 页码:1063-1083
关键词	Multi-target tracking Multi-dimensional assignment Rank-1 tensor approximation Data association
ISSN号	0920-5691
DOI	10.1007/s11263-018-01147-z
通讯作者	Hu, Weiming(wmhu@nlpr.ia.ac.cn)
英文摘要	High-order motion information is important in multi-target tracking (MTT) especially when dealing with large inter-target ambiguities. Such high-order information can be naturally modeled as a multi-dimensional assignment (MDA) problem, whose global solution is however intractable in general. In this paper, we propose a novel framework to the problem by reshaping MTT as a rank-1 tensor approximation problem (R1TA). We first show that MDA and R1TA share the same objective function and similar constraints. This discovery opens a door to use high-order tensor analysis for MTT and suggests the exploration of R1TA. In particular, we develop a tensor power iteration algorithm to effectively capture high-order motion information as well as appearance variation. The proposed algorithm is evaluated on a diverse set of datasets including aerial video sequences containing ariel borne dense highway scenes, top-view pedestrian trajectories, multiple similar objects, normal view pedestrians and vehicles. The effectiveness of the proposed algorithm is clearly demonstrated in these experiments.
资助项目	Beijing Natural Science Foundation[L172051] ; Natural Science Foundation of China[61502492] ; Natural Science Foundation of China[61751212] ; Natural Science Foundation of China[61721004] ; NSFC-general technology collaborative Fund for basic research[U1636218] ; Key Research Program of Frontier Sciences, CAS[QYZDJ-SSW-JSC040] ; CAS External cooperation key project ; US NSF[1814745] ; US NSF[1407156] ; US NSF[1350521]
WOS关键词	ALGORITHM
WOS研究方向	Computer Science
语种	英语
出版者	SPRINGER
WOS记录号	WOS:000474559000006
资助机构	Beijing Natural Science Foundation ; Natural Science Foundation of China ; NSFC-general technology collaborative Fund for basic research ; Key Research Program of Frontier Sciences, CAS ; CAS External cooperation key project ; US NSF
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/26849]
专题	中国科学院自动化研究所
通讯作者	Hu, Weiming
作者单位	1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing, Peoples R China 2.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, CAS Ctr Excellence Brain Sci & Intelligence Techn, Beijing, Peoples R China 3.Univ Chinese Acad Sci, Beijing, Peoples R China 4.Temple Univ, Dept Comp & Informat Sci, Philadelphia, PA 19122 USA
推荐引用方式 GB/T 7714	Shi, Xinchu,Ling, Haibin,Pang, Yu,et al. Rank-1 Tensor Approximation for High-Order Association in Multi-target Tracking[J]. INTERNATIONAL JOURNAL OF COMPUTER VISION,2019,127(8):1063-1083.
APA	Shi, Xinchu,Ling, Haibin,Pang, Yu,Hu, Weiming,Chu, Peng,&Xing, Junliang.(2019).Rank-1 Tensor Approximation for High-Order Association in Multi-target Tracking.INTERNATIONAL JOURNAL OF COMPUTER VISION,127(8),1063-1083.
MLA	Shi, Xinchu,et al."Rank-1 Tensor Approximation for High-Order Association in Multi-target Tracking".INTERNATIONAL JOURNAL OF COMPUTER VISION 127.8(2019):1063-1083.