Abnormal Activity Recognition on Surveillance: A Review

Abstract

Utilizing surveillance cameras to monitor public behaviour has become more important recently for public safety at various sites due to a rise in crimes. Maintaining safety and security has become a survival issue for people due to rising crime rates. With the advancement of security cameras, they now serve as a constant watch on public behavior. Many current surveillance systems require a human operator to continuously monitor them since the amount of video data is increasing daily, making them ineffective. To automatically identify odd behaviour in both public and private places, modern surveillance systems must be intelligent. In current decade, due to industrial 4.0 revolution, machine learning and deep learning based intelligent algorithms plays major role in providing efficient performance in various applications exclusively over automatic detection and identification. This research article mainly focuses on investigation of different intelligent algorithms used for an effective recognition of abnormal activity through surveillance, its major advantages, challenges and contributions in terms of various applications.