Suspicious Activity Detection using Image Processing

Abstract

Nowadays, video surveillance is used almost everywhere for security. The traditional method of monitoring cameras requires constant human intervention. Using Deep Learning and Image Processing, the proposed work aims to eliminate time and effort wasted on monitoring video surveillance cameras. Predicting human behavior is almost impossible. Deep Learning is used to detect suspicious and non-suspicious activity and to warn the user if any suspicious activity is detected. The proposed system strives for the detection of real-world suspicious activities such as burglaries, assaults etc. in surveillance videos.