Recent advancement in anomaly detection in surveillance videos

Abstract

Detection of Anomaly is of a notable and emergent problem into many diverse fields like information theory, deep learning, computer vision, machine learning, and statistics that have been researched within the various application from diverse domains including agriculture, health care, banking, education, and transport anomaly detection. Newly, numbers of important anomaly detection techniques along with diverseness of sort have been watched. The main aim of this paper to come up with a broad summary of the present development on detection of an anomaly, exclusively for video data with mixed types and high dimensionalities, where identifying the anomalous behaviors and event or anomalous patterns is a significant task. The paper expresses the advantages and disadvantages of the detection methods the experiments tried on the publically available benchmark dataset to assess numerous popular and classical methods and models. The objective of this analysis is to furnish an understanding of recent computer vision and machine algorithms methods and also state-of-the-art deep learnings techniques to detect anomalies for researchers. At last, the paper delivered roughly directions for future research on an anomalies detection.