Video Anomaly Detection Based on Mixed Statistic Feature
Hao Yu 1, Xinghao Jiang 1,3,
Tanfeng Sun 1,2,3, and
Shilin Wang 1,3
1. School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University
2. State Key Laboratory of Software Engineering, Wuhan University, China
3. National Engineering Lab on Information Content Analysis Techniques, GT036001, Shanghai, China
2. State Key Laboratory of Software Engineering, Wuhan University, China
3. National Engineering Lab on Information Content Analysis Techniques, GT036001, Shanghai, China
Abstract—In this paper, a new efficient method based on mixed statistic feature is proposed for video anomaly detection in densely crowded scenes. The proposed mixed statistic feature is a hand-designed feature considering both magnitude and phase information of the optical flow block after the preprocessing step which based on the latent consistency information of moving objects in the block. Gaussian Mixture Model (GMM) is employed in our method to establish the appropriate probability model for our block-based feature. Experimental results on the challenging UCSD datasets (Ped1 and Ped2) have shown that our method outperformed four state-of-the-art approaches both in accuracy and efficiency (less than 1s per frame in Matlab environment).
Index Terms—mixed statistic feature, block-based, GMM, anomaly detection, crowded scenes
Cite: Hao Yu, Xinghao Jiang, Tanfeng Sun, and Shilin Wang, "Video Anomaly Detection Based on Mixed Statistic Feature," International Journal of Signal Processing Systems, Vol. 4, No. 2, pp. 150-154, April 2016. doi: 10.12720/ijsps.4.2.150-154
Cite: Hao Yu, Xinghao Jiang, Tanfeng Sun, and Shilin Wang, "Video Anomaly Detection Based on Mixed Statistic Feature," International Journal of Signal Processing Systems, Vol. 4, No. 2, pp. 150-154, April 2016. doi: 10.12720/ijsps.4.2.150-154