Multitarget Tracking in Nonoverlapping Cameras Using a Reference Set

被引:20
作者
Chen, Xiaojing [1 ]
An, Le [2 ]
Bhanu, Bir [3 ]
机构
[1] Univ Calif Riverside, Dept Comp Sci, Riverside, CA 92521 USA
[2] Univ N Carolina, Biomed Res Imaging Ctr, Chapel Hill, NC 27599 USA
[3] Univ Calif Riverside, Ctr Res Intelligent Syst, Riverside, CA 92521 USA
基金
美国国家科学基金会;
关键词
Multi-target tracking; reference set; surveillance; OBJECT TRACKING; MODEL; SURVEILLANCE; RECOGNITION; APPEARANCE;
D O I
10.1109/JSEN.2015.2392781
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Tracking multiple targets in nonoverlapping cameras are challenging since the observations of the same targets are often separated by time and space. There might be significant appearance change of a target across camera views caused by variations in illumination conditions, poses, and camera imaging characteristics. Consequently, the same target may appear very different in two cameras. Therefore, associating tracks in different camera views directly based on their appearance similarity is difficult and prone to error. In most previous methods, the appearance similarity is computed either using color histograms or based on pretrained brightness transfer function that maps color between cameras. In this paper, a novel reference set based appearance model is proposed to improve multitarget tracking in a network of nonoverlapping cameras. Contrary to previous work, a reference set is constructed for a pair of cameras, containing subjects appearing in both camera views. For track association, instead of directly comparing the appearance of two targets in different camera views, they are compared indirectly via the reference set. Besides global color histograms, texture and shape features are extracted at different locations of a target, and AdaBoost is used to learn the discriminative power of each feature. The effectiveness of the proposed method over the state of the art on two challenging real-world multicamera video data sets is demonstrated by thorough experiments.
引用
收藏
页码:2692 / 2704
页数:13
相关论文
共 46 条
[1]  
An L, 2013, 2013 10TH IEEE INTERNATIONAL CONFERENCE ON ADVANCED VIDEO AND SIGNAL BASED SURVEILLANCE (AVSS 2013), P244, DOI 10.1109/AVSS.2013.6636647
[2]   Dynamic Bayesian Network for Unconstrained Face Recognition in Surveillance Camera Networks [J].
An, Le ;
Kafai, Mehran ;
Bhanu, Bir .
IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS, 2013, 3 (02) :155-164
[3]  
[Anonymous], 2012, Proceedings of IEEE International Conference on Distributed Smart Cameras (ICDSC)
[4]  
[Anonymous], 2013, P 29 ANN AC M POW SY
[5]  
Black J, 2002, IEEE WORKSHOP ON MOTION AND VIDEO COMPUTING (MOTION 2002), PROCEEDINGS, P169, DOI 10.1109/MOTION.2002.1182230
[6]   Understanding Transit Scenes: A Survey on Human Behavior-Recognition Algorithms [J].
Candamo, Joshua ;
Shreve, Matthew ;
Goldgof, Dmitry B. ;
Sapper, Deborah B. ;
Kasturi, Rangachar .
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 2010, 11 (01) :206-224
[7]  
Chang TH, 2001, 2001 IEEE WORKSHOP ON MULTI-OBJECT TRACKING, PROCEEDINGS, P19, DOI 10.1109/MOT.2001.937977
[8]   An Online Learned Elementary Grouping Model for Multi-target Tracking [J].
Chen, Xiaojing ;
Qin, Zhen ;
An, Le ;
Bhanu, Bir .
2014 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2014, :1242-1249
[9]   Object tracking across non-overlapping views by learning inter-camera transfer models [J].
Chen, Xiaotang ;
Huang, Kaiqi ;
Tan, Tieniu .
PATTERN RECOGNITION, 2014, 47 (03) :1126-1137
[10]   Two-Way Full-Duplex Amplify-and-Forward Relaying [J].
Cheng, Xilin ;
Yu, Bo ;
Cheng, Xiang ;
Yang, Liuqing .
2013 IEEE MILITARY COMMUNICATIONS CONFERENCE (MILCOM 2013), 2013, :1-6