Real Time Human Activity Detection using YOLOv7

Abstract

The You Only Look Once (YOLO) algorithm is used in this research to demonstrate a sophisticated human activity detection system integrated with real-time applications. Computer vision, a subset of deep learning, is crucial in grasping the content of pictures and videos and having applications in security monitoring, healthcare, traffic control, and other fields. The suggested system focuses on three critical safety scenarios, fall detection, social distance detection, and drowsiness detection, to improve real-time safety, security, and efficiency. The system uses YOLO's extraordinary strength of object identification and localization to deliver real-time processing, high accuracy, and flexibility to dynamic situations. The modules' accuracy and dependability are established through extensive testing and assessment, making them useful instruments for addressing challenges in the real world.