Optimizing Laboratory Safety Compliance Using Computer Vision: Real-Time PPE Detection in Laminar Flow Cabinets

This research focuses on enhancing laboratory safety through an AI-powered Personal Protective Equipment (PPE) detection system using a customized YOLOv8 model. With current safety protocols requiring robust monitoring of PPE usage, the aim was to develop a scalable model for real-time compliance detection. Methodologically, a dataset of 2367 images was employed, trained, and validated on Roboflow servers, employing SGD with tailored parameters. Results indicate a 95% accuracy in PPE detection, with a 98.4% detection rate and 99.9% false alarm rate, showcasing high operational efficiency. Integrating survey data furthered insights into real-world PPE usage patterns, crucial for deployment. Future efforts will focus on refining model limitations, expanding applications to diverse laboratory scenarios, and integrating user behavior analysis to bolster safety protocols comprehensively.