Please wait a minute...
Frontiers of Computer Science

ISSN 2095-2228

ISSN 2095-2236(Online)

CN 10-1014/TP

Postal Subscription Code 80-970

2018 Impact Factor: 1.129

Front Comput Sci Chin    2011, Vol. 5 Issue (3) : 279-289    https://doi.org/10.1007/s11704-011-0080-4
RESEARCH ARTICLE
Human behavior clustering for anomaly detection
Xudong ZHU(), Zhijing LIU
School of Computer Science and Technology, Xidian University, Xi’an 710071, China
 Download: PDF(413 KB)   HTML
 Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks
Abstract

This paper aims to address the problem of modeling human behavior patterns captured in surveillance videos for the application of online normal behavior recognition and anomaly detection. A novel framework is developed for automatic behavior modeling and online anomaly detection without the need for manual labeling of the training data set. The framework consists of the following key components. 1) A compact and effective behavior representation method is developed based on spatial-temporal interest point detection. 2) The natural grouping of behavior patterns is determined through a novel clustering algorithm, topic hidden Markov model (THMM) built upon the existing hidden Markov model (HMM) and latent Dirichlet allocation (LDA), which overcomes the current limitations in accuracy, robustness, and computational efficiency. The new model is a four-level hierarchical Bayesian model, in which each video is modeled as a Markov chain of behavior patterns where each behavior pattern is a distribution over some segments of the video. Each of these segments in the video can be modeled as a mixture of actions where each action is a distribution over spatial-temporal words. 3) An online anomaly measure is introduced to detect abnormal behavior, whereas normal behavior is recognized by runtime accumulative visual evidence using the likelihood ratio test (LRT) method. Experimental results demonstrate the effectiveness and robustness of our approach using noisy and sparse data sets collected from a real surveillance scenario.

Keywords computer vision      unsupervised anomaly detection      Bayesian topic models      hidden Markov model (HMM)      spatiotemporal interest points     
Corresponding Author(s): ZHU Xudong,Email:zhudongxu@vip.sina.com   
Issue Date: 05 September 2011
 Cite this article:   
Xudong ZHU,Zhijing LIU. Human behavior clustering for anomaly detection[J]. Front Comput Sci Chin, 2011, 5(3): 279-289.
 URL:  
https://academic.hep.com.cn/fcs/EN/10.1007/s11704-011-0080-4
https://academic.hep.com.cn/fcs/EN/Y2011/V5/I3/279
Fig.1  Topic hidden Markov model
Fig.2  (a) Example frames from video sequences in our scene; (b) example topics/actions learned in our scene, illustrated by the most likely spatial-temporal feature points; (c) behavior patterns discovered in our scene, illustrated by the most likely spatial-temporal feature points for each behavior and the transitions between behaviors
Fig.3  Example of anomaly detection in the entrance/exit area of an office building. (a)–(d) An abnormal behavior where a person attempts to scratch the car parked in the area. It resembles C3 in the early stages; (e) behavior was detected as an anomaly from clip 62 until the end based on
Fig.4  (a) Performance of anomaly detection using our behavior models trained by unlabeled data. The mean ROC curve was obtained by varying ; (b) performance of behavior recognition using models trained by unlabeled data
MethodsAccuracy/%
Topic HMM89.26
HMM (Xiang et al. [20])82.76
HTMM (Gruber et al. [9])82.46
n-grams (Hamid et al. [19])82.38
LDA (Niebles et al. [6])81.50
MAP-based (Boiman et al. [10])79.32
LSA (Zhong et al. [5])76.56
Tab.1  Comparison of different methods.
1 Dollar P, Rabaud V, Cottrell G, Belongie S. Behavior recognition via sparse spatio-temporal features. In: Proceedings of 14th International Conference on Computer Communications and Networks . 2005, 65–72
2 Yilmaz A. Shah M. Recognizing human actions in videos acquired by uncalibrated moving cameras. In: Proceedings of 10th IEEE International Conference on Computer Vision . 2005, 150–157
3 Song Y, Goncalves L, Perona P. Unsupervised learning of human motion. IEEE Transactions on Pattern Analysis and Machine Intelligence , 2003, 25(7): 814–827
doi: 10.1109/TPAMI.2003.1206511
4 Fanti C, Zelnik-Manor L, Perona P. Hybrid models for human motion recognition. In: Proceedings of 10th IEEE International Conference on Computer Vision . 2005, 1166–1173
5 Zhong H, Shi J, Visontai M. Detecting unusual activity in video. In: Proceedings of 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition . 2004, 819–826
doi: 10.1109/CVPR.2004.1315249
6 Niebles J C, Wang H C, Li F F. Unsupervised learning of human action categories using spatial-temporal words. International Journal of Computer Vision , 2008, 79(3): 299–318
doi: 10.1007/s11263-007-0122-4
7 Wallach H M. Topic modeling: beyond bag-of-words. In: Proceedings of 23rd International Conference on Machine Learning . 2006, 977–984
doi: 10.1145/1143844.1143967
8 Wang X, McCallum A. A note on topical n-grams. Technical Report UM-CS-071. Department of Computer Science University of Massachusetts Amherst , 2005
9 Gruber A, Rosen-Zvi M, Weiss Y. Hidden topic Markov models. In: Proceedings of Artificial Intelligence and Statistics . 2007
10 Boiman O, Irani M. Detecting irregularities in images and in video. In: Proceedings of 10th IEEE International Conference on Computer Vision . 2005, 462–469
11 Oliver N M, Rosario B, Pentland A P. ABayesian computer vision system for modeling human interactions. IEEE Transactions on Pattern Analysis and Machine Intelligence , 2000, 22(8): 831–843
doi: 10.1109/34.868684
12 Zelnik-Manor L, Irani M. Event-based analysis of video. In: Proceedings of 2001 IEEE Conference on Computer Vision and Pattern Recognition . 2001, 123–130
13 Hongeng S. Nevatia R. Multi-agent event recognition. In: Proceedings of 8th IEEE International Conference on Computer Vision . 2001, 84–91
14 Russo R, Shah M, Lobo N. A computer vision system for monitoring production of fast food. In: Proceedings of 5th Asian Conference on Computer Vision, Melbourne . 2002, 23–25
15 Johnson N, Hogg D. Learning the distribution of object trajectories for event recognition. Image and Vision Computing , 1995, 14(8): 609–615
16 Brand M, Oliver N, Pentland A. Coupled hidden Markov models for complex action recognition. In: Proceedings of 1997 IEEE Computer Society Conference on Computer Vision and Pattern Recognition . 1997, 994–999
doi: 10.1109/CVPR.1997.609450
17 Medioni G, Cohen I, Bremond F, Hongeng S, Nevatia R. Event detection and analysis from video streams. IEEE Transactions on Pattern Analysis and Machine Intelligence , 2001, 23(8): 873–889
doi: 10.1109/34.946990
18 Naphide H R, Huang T S. A probabilistic framework for semantic video indexing, filtering, and retrieval. IEEE Transactions on Multimedia , 2001, 3(1): 141–151
doi: 10.1109/6046.909601
19 Hamid R, Johnson A, Batta S, Bobick A, Isbell C, Coleman G. Detection and explanation of anomalous activities: representing activities as bags of event n-grams. In: Proceedings of 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition . 2005, 1031–1038
20 Xiang T, Gong S. Beyond tracking: modelling activity and understanding behaviour. International Journal of Computer Vision , 2006, 67(1): 21–51
doi: 10.1007/s11263-006-4329-6
21 Wilpon J G, Rabiner L R, Lee C H, Goldman E R. Automatic recognition of keywords in unconstrained speech using hidden Markov models. IEEE Transactions on Acoustics, Speech, and Signal Processing , 1990, 38(11): 1870–1878
doi: 10.1109/29.103088
22 Blei D M, Ng A Y, Jordan M I, Lafferty J. Latent Dirichlet allocation. Journal of Machine Learning Research , 2003, 3: 993–1022
doi: 10.1162/jmlr.2003.3.4-5.993
23 Griffiths T L, Steyvers M. Finding scientific topic. Proceedings of the National Academy of Sciences of the United States of America. 2004, 5228–5235
[1] Ahmad ALI,Abdul JALIL,Jianwei NIU,Xiaoke ZHAO,Saima RATHORE,Javed AHMED,Muhammad AKSAM IFTIKHAR. Visual object tracking—classical and contemporary approaches[J]. Front. Comput. Sci., 2016, 10(1): 167-188.
[2] Hui LI,Yun LIU,Shengwu XIONG,Lin WANG. Pedestrian detection algorithm based on video sequences and laser point cloud[J]. Front. Comput. Sci., 2015, 9(3): 402-414.
[3] Wai-Ki CHING, Ho-Yin LEUNG, Zhenyu WU, Hao JIANG. Modeling default risk via a hidden Markov model of multiple sequences[J]. Front Comput Sci Chin, 2010, 4(2): 187-195.
[4] WU Wenjun, Wen-Tsun Wu, GAO Xiaoshan. Mathematics mechanization and applications after thirty years[J]. Front. Comput. Sci., 2007, 1(1): 1-8.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed