Detection of Suspicious Activities in a Surveillance Video using convolution neural network

Abstract

Detection of suspicious human activities in automated video surveillance applications, is of great practical importance in real world. Reliable classification of suspicious human movements can be very difficult due to the random nature of human movements. The primary aim of the project is to define an approach to the problem of automatically tracking people and detecting unusual or suspicious movements in CCTV videos. Firstly, the videos are converted into frames of specified size. Then from the obtained frames, humans and human actions are detected from the video using a background subtraction method. Then the features are extracted using a convolutional neural network (CNN). The features thus extracted are fed to a Trained CNN model. Labelled videos of some suspicious activities are also fed to the CNN and their features are also extracted. Then the features extracted using Convolutional Neural Network (CNN) are compared against these features extracted from the labelled sample video of classified suspicious actions using a Trained model  and various suspicious activities are detected from the given video.