Abstract

In this paper, an artificial intelligence based video monitoring system for security applications is being presented. The inspiration driving this undertaking work is to enhance the present security frameworks. The present system is human checking which has to be eradicated by developing automated tools. The present security administrations depend a great deal on closed circuit television (CCTV) cameras and video filming. This innovation has altered security segments with innovative results. The checking administrations require steady observing by an individual. It is squandering human resources and is inclined to make mistakes. These imperfections leave the framework powerless. The proposed arrangement gives a self-sufficient framework to screen the previously mentioned exercises. The framework will have the option to consistently screen the video which is taken. The principle thought is to utilize Graphics Processing Unit (GPU) quickened Deep Learning techniques to prepare a convolutional neural network (CNN) to distinguish ill-conceived acts. The work done in this paper shows the effectiveness of the method that is being proposed.