Authors: Assistant Professor Dr.B.Nageshwar Rao, K. Sushmitha
Abstract: The increasing demand for intelligent surveillance systems has accelerated the adoption of artificial intelligence and deep learning technologies for improving public safety and crime prevention. monitoring, making them susceptible to delayed responses, operator fatigue, and human error, particularly in crowded public environments. Detecting violent incidents in real time remains a significant challenge because aggressive activities often occur unexpectedly and require immediate intervention. automated video analysis systems capable of identifying complex human activities with high accuracy, thereby supporting proactive surveillance and timely emergency response. Lightweight convolutional neural network architectures and temporal sequence learning models have further in real-world surveillance environments. This study presents an intelligent videos by integrating MobileNet and The proposed framework is evaluated using standard classification metrics including Accuracy, Precision, Recall, and F1-Score. Experimental results demonstrate that the integrated MobileNet–BiLSTM model achieves an overall classification accuracy of approximately 96%, with balanced precision and recall for both violent and non-violent activity recognition. The lightweight architecture significantly reduces computational complexity while maintaining high prediction performance, making the framework suitable for deployment on resource-constrained surveillance devices. The developed system enables rapid identification of violent incidents, supports faster emergency response by law enforcement agencies, and provides a scalable solution for intelligent public safety monitoring in transportation hubs, educational institutions, commercial facilities, and other high-risk public environments.
