A Comprehensive Study on Automated Anomaly Detection Techniques in Video Surveillance

Abstract

Video surveillance systems are the most important aspect of security systems which are having wide range of applications in ourevery day life. In particular, it plays a significant role in remote monitoring of facilities in public and private premises. In this context, video surveillance refers to observing the scenes of improper human behaviours which are termed as real world anomalies. But, the traditional way of involving humans for real world anomaly detection isa time consuming process and involves various overheads. Thus, an automated anomaly detection in video surveillance using intelligent methods becomes an important area of research. This paper gives a comprehensive study on automated anomaly detection in video surveillance based on statistical, proximity, classification, reconstruction and prediction approaches with a special focus towards crime detection. In addition, this paper also highlights various benchmark datasets used in automated anomalydetection in video surveillance.