A Novel Deep Learning-, Camera-, and Sensor-Based System for Enforcing Hand Hygiene Compliance in Healthcare Facilities

Abstract

Hospital-acquired infections (HAIs) are a major cause of death worldwide, and poor hand hygiene compliance is a primary reason for their spread. This article proposes an artificial intelligence-, microcontroller-, and sensor-based system that monitors and improves staff hand hygiene compliance at various critical points in a hospital. The system uses a convolutional neural network (CNN) to detect and track if staff have followed the World Health Organization (WHO) handrub/handwash guidelines at alcohol dispensers, hospital sinks, and patient beds. The system also uses radio frequency identification (RFID) tags, vibration motors, LEDs, and a central server to identify staff, alert them of their cleaning requirements, monitor their cleaning activity, and report compliance data. We obtain an accuracy of 90.6% in classifying all steps of the WHO-stipulated handwash/handrub guidelines and a testing accuracy of 89.8% on Ivanovs et al.’s dataset. The system ensures that hospital staff stay compliant with all WHO hand hygiene guidelines, saving countless lives.