Author : Shreekar Waghela 1
Date of Publication :3rd August 2017
Abstract: Video surveillance is very essential as threat and crime increases. In this project, a method is proposed in which the behaviour of crowd is detected without individual tracking of objects in a frame. This method is based on the motion intensity of the crowd which can be determined by accumulating all optical flow vectors of a frame. The abnormal crowd activity can then be detected by setting up a threshold to detect any sudden change in motion intensity.
Reference :
-
[1] H. Mousavi, S. Mohammadi, Alessandro Perina, R. Chellali, Vittorio Murino,“Analyzing Tracklets for the Detection of Abnormal Crowd Behaviour”, IEEE Winter Conference on Applications of Computer Vision, 2015.
[2] B.Yogameena, K.S. Priya, “ Synoptic Video Based Human Crowd Behaviour Analysis for Forensic Video Surveillance”, IEEE International Conference on Advances in Pattern Recognition(ICAPR), 2015.
[3] Aravinda Rao, J. Gubbi, M. Slaven , M. Palaniswami, “Crowd Event Detection on Optical Flow Manifolds”, IEEE transactions on cybernetics, 2016
[4] Yang , Xia Li and Limin Jia, “Abnormal Crowd Behavior Detection Based on Optical Flow and Dynamic Threshold”, IEEE International Conference on Intelligent Control and Automation, 2014.
[5] B. Solmaz, Brian E. Moore and Mubarak Shah, “Identifying Behaviors in Crowd Scenes Using Stability Analysis for Dynamical Systems”, IEEE Transactions on pattern analysis and machine intelligene, 2010.
[6] Guo Xiong, Xinyu Wu, Yen-Lun Chen and Yong Ou, “Abnormal Crowd Behavior Detection Based on the Energy Model”, IEEE International Conference on Information and Automation, 2011.
[7] Zhi Zhong, Weizhong Ye, “Crowd Energy and Feature Analysis”, IEEE International Conference on Integration Technology, 2007.