Online learning affinity measure with CovBoost for multi-target tracking

doi:10.1016/j.neucom.2015.05.093

	Online learning affinity measure with CovBoost for multi-target tracking
	Li, Guorong 1,3; Huang, Qingming 1,2,3; Jiang, Shuqiang 2; Xu, Yingkun 2; Zhang, Weigang 4
	2015-11-30
发表期刊	NEUROCOMPUTING
ISSN	0925-2312
卷号	168 页码:327-335
摘要	In this paper, we propose a new online learning method for measuring affinity between tracklets in multi-target tracking. As targets and background usually keep changing in the video, fixed affinity measurement could not adapt to their variations. Most existing affinity learning methods construct labeled samples based on the obtained tracklets, and then minimize a predefined loss function to get an optimal affinity measurement. However, those methods simply assume that the training error equals to testing error which is not true in many of real time tracking scenarios. Differently, we propose to learn affinity measurement through CovBoosting, which considers the evolution of the tracklets and could obtain affinity measurement with more discriminative ability. To deal with targets' disappearance and new targets' appearance, we combine tracklet affinity with contextual information to do an optimal inference. Moreover, an online updating algorithm is developed to guarantee that the learned tracklet affinity is always optimal for tracking targets in current sliding window. Experimental results on benchmark datasets demonstrate that tracklet affinity learned with our method is more discriminative and could greatly improve the performance of the multi-target tracker. (C) 2015 Elsevier B.V. All rights reserved.
关键词	Multi-target tracking Tracklet affinity CovBoost Online learning
DOI	10.1016/j.neucom.2015.05.093
收录类别	SCI
语种	英语
资助项目	National Basic Research Program of China (973 Program)[2012CB316400] ; National Natural Science Foundation of China[61303153] ; National Natural Science Foundation of China[61025011] ; National Natural Science Foundation of China[61332016] ; National Natural Science Foundation of China[61202322] ; China Postdoctoral Science Foundation[2014T70111] ; President Foundation of UCAS
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000359165000033
出版者	ELSEVIER SCIENCE BV
引用统计	被引频次：4[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/9494
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Jiang, Shuqiang
作者单位	1.Univ Chinese Acad Sci CAS, Beijing 100190, Peoples R China 2.Chinese Acad Sci, Inst Comput Tech, Key Lab Intell Info Proc, Beijing 100080, Peoples R China 3.Chinese Acad Sci, Key Lab Big Data Min & Knowledge Management, Beijing, Peoples R China 4.Harbin Inst Tech, Sch Comp Sci & Technol, Harbin, Heilongjiang, Peoples R China
推荐引用方式 GB/T 7714	Li, Guorong,Huang, Qingming,Jiang, Shuqiang,et al. Online learning affinity measure with CovBoost for multi-target tracking[J]. NEUROCOMPUTING,2015,168:327-335.
APA	Li, Guorong,Huang, Qingming,Jiang, Shuqiang,Xu, Yingkun,&Zhang, Weigang.(2015).Online learning affinity measure with CovBoost for multi-target tracking.NEUROCOMPUTING,168,327-335.
MLA	Li, Guorong,et al."Online learning affinity measure with CovBoost for multi-target tracking".NEUROCOMPUTING 168(2015):327-335.