A Review on Crowd Behavior Analysis Methods for Video Surveillance

Abstract

Abnormal Crowd Detection has become the most viral and active research topic in computer vision. There is need of automated tracking of the abnormalities in surveillance video sequence for the detection of abnormal events. These systems are mainly used for supervising and security purpose. This automated system can alarm for abnormality in fairs, temples. It can also be used for the traffic monitoring etc. Crowd is a group of individuals belonging to a community or society. In the crowd, there exist many behavior abnormalities. Crowd density estimation, crowd motion detection, crowd tracking and crowd behavior recognition are multiple techniques for detecting abnormalities. Computer based crowd analysis algorithm can be divided into three groups; people counting, people tracking and crowd behavior analysis. In this paper, we will discuss multiple techniques of abnormal crowd detection background subtraction, optical flow, 3D Convolutional neural network, hydrodynamics lens. Once detected, a moving object could be classified as a human being using shape-based, texture-based or motion-based features. Comparison of available techniques for detecting abnormal crowd in surveillance videos has also been done in this paper.