Sentinel Eyes Violence Detection System

In recent years, the proliferation of surveillance systems has increased the demand for effective methods to detect violent activities in various environments. This project proposes a comprehensive approach to violence detection by integrating state-of-the-art computer vision and deep learning techniques. This study uses YOLOv8, OpenPose, and LSTM networks to present a multi-modal technique for violence detection. Real-time object detection using YOLOv8 is done with an emphasis on human identification. OpenPose gathers comprehensive data on human posture, and LSTM networks use temporal pattern analysis to identify violent behavior. This platform uses OpenPose to coordinate multi-person 2D plan forecasts, Yolov8 to quickly locate individuals, and a combination of CNN and long short-term memory (LSTM) to classify harmful conduct. By integrating these elements, a strong violence detection system that incorporates temporal and spatial awareness is intended to be created. Benchmark datasets will be used to assess the project's efficacy, with possible surveillance and public safety uses.