Real Time Unusual Event Detection in Video Sequences

Abstract: Detection of human activities accurately in the video surveillance system is very difficult task. For diverse application areas, while detecting human activities, basically three aspects are taking in concern for human activity recognition system: segmenting object of interest, feature extraction & representation and human activity detection & classification. In this paper we have proposed a basic framework for detection suspicious human behaviour as well as tracking of detected human who is doing some unusual activity. The purpose of these methods could lead to better understanding to the suspicious activity of single person as well as multiple persons. The improved design of built environment will provide better surveillance. Such system can be implemented at the public places such as shopping malls, airports, and railway-station or any private premises where security is the prime most concern.

Keywords: Video surveillance, Human activity detection & classification, Unusual event, Skeletonization, Thinning.