DEEPGUARD: Enhancing Violence Detection in Smart Cities Through Deep Learning

This project, "DeepGuard: Enhancing Violence Detection in Smart Cities Through Deep Learning," focuses on improving surveillance analysis by detecting and classifying suspicious activities in video footage using advanced computer vision and deep learning models.The system utilizes the SlowFast model for activity classification, which processes both slow and fast video frames to capture detailed temporal dynamics, enabling the identification of complex and subtle human behaviors.To enhance detection accuracy, additional models are employed to classify activities based on visual patterns and context, improving the system's capability to differentiate between normal and abnormal behavior.YOLOv5 is used for detecting specific security threats such as weapons, accidents, and explosions in recorded video data.To identify physical confrontations, the system analyzes body movements to detect fighting behavior.By integrating these deep learning techniques, the proposed system offers a comprehensive framework for post-event video analysis in smart city environments.It supports the monitoring of public and private areas by helping authorities review and assess potential threats effectively, thereby contributing to improved urban safety and situational awareness.