Contactless Hand Sanitizer System with Machine Learning Verification in Reducing Healthcare-Associated Infection (HAIs) - An Initial Study

Healthcare-associated infections (HAIs) remain a huge concern in most healthcare facilities, mainly caused by the inability to perform proper hand hygiene and poor compliance with established hand hygiene practices. Behavioural lapses in hand hygiene significantly contribute to the transmission of pathogens and the persistence of Healthcare-associated infections (HAIs) within clinical environments. HAIs remain a huge concern in most healthcare facilities, mainly caused by the inability to perform proper hand hygiene. The study designs a contactless hand sanitizer that incorporates an automated access control for ensuring proper hygiene and limiting cross-contamination in areas that require high sterility, such as the Intensive Care Unit (ICU). The system, built using Raspberry Pi 4 and components like ultrasonic sensors, IR sensors, a UV light, and the OV5647 camera, dispenses sanitizer and verifies compliance before unlocking the door. A Convolution Neural Network (CNN), MobileNetV2, was trained on ultraviolet (UV)-lit images of sanitized and unsanitized hands to detect the presence of fluorescent residue. It analyses the presence of fluorescent liquid in hand sanitizer for compliance before granting access. While the model demonstrated high accuracy during training, hardware limitations, especially the camera’s low sensitivity under UV light, affected its real-time performance. Nevertheless, the system provides an initial basis that exemplifies the potential of machine learning-integrated sanitary enforcement as an initial point of further development in the direction of more comprehensive approaches to reducing HAIs.