Public Space Real-Time Violence Detection and Notifier

This paper introduces a real time system for detecting violence using Deep Learning and Convolutional Neural Networks (CNN). The increasing need for security in areas calls for effective ways to quickly identify violent behavior. We provide a real-time video flow evaluation system that makes use of CNNs to fast identify violent behaviors. Our method reduces fake positives and improves accuracy over current techniques. Robust overall performance in various settings is ensured through the modules for preprocessing, feature extraction, classification, and real-time detection, ensuring robust performance in diverse scenarios. We provide a detailed algorithm and methodology for implementing the system, highlighting its potential for future enhancements.