Safety Watch : A Fusion Approach of YOLOv3 Object Detection and Convolutional Neural Network for Real-Time Motorcycle Helmet Compliance Monitoring

This project introduces a novel system for enhancing motorcycle safety through automated helmet detection.By integrating YOLOv3 for motorcycle identification and a Convolutional Neural Network (CNN) for helmet classification, the system offers real-time monitoring of helmet compliance.A graphical user interface (GUI) built with Tkinter facilitates user interaction, featuring a login window for authentication and video upload functionality.Upon login, users can upload videos for processing.The system employs YOLOv3 to detect motorcycles within the footage, extracting regions of interest (ROIs) encompassing riders' heads.These ROIs undergo scrutiny by the CNN model to ascertain helmet presence, with detected helmets annotated onto the video frames.The intuitive GUI streamlines user engagement, enhancing accessibility and usability.This integrated approach promotes proactive motorcycle safety measures, empowering stakeholders to monitor and enforce helmet compliance effectively.The system's automated nature ensures timely intervention, reducing the risk of head injuries and fostering safer road practices among motorcyclists.Additionally, robust user authentication mechanisms safeguard data integrity and confidentiality, bolstering the system's reliability and trustworthiness.