An Automatic Violence Detection Technique Using 3D Convolutional Neural Network

In the field of machine learning, the deep learning technique plays a very important role as it is useful in various real-life domains. As various crimes and misdeeds are occurring in various public places because of lack of proper monitoring, a number of methods have been proposed for detecting violence from videos. Automatic violence detection has gained increased research importance in case of video surveillance. However, they suffer from various limitations and most of the times it depends on special criteria. In this perspective, this paper proposes an effective violence detection method from videos using 3D convolutional neural network. The proposed methodology uses machine learning and deep learning techniques for improving the accuracy. Comprehensive performance analyses have proven that the proposed method achieves high performance in case of detecting violence from videos. The experimental results also prove that the proposed technique outperforms various other existing methods for detecting violence from videos.