A Review of Deep Learning Algorithms for Anomaly Detection in Videos

Abstract

Anomaly detection in video assists in the resolution of a wide range of problems. A robust anomaly detection model is necessary due to the growing use of surveillance cameras in both indoor and outdoor settings. As a result, numerous strategies have been proposed in this field. Anomaly detection has already been the subject of several surveys. This survey focuses on deep learning approaches based on video anomaly detection. we categorize the various Deep Learning approaches according to their objectives like score based, future frame-based, Clasiification and reconstruction error based approaches. Additionally, it discusses evaluation criteria and commonly used datasets. We also suggest some possible directions future directions for research.