Embedded System for Automatic Mask Detection using YOLOv4 Deep Learning and PyQt5 Interface

The use of masks remains crucial, especially in high-risk areas for disease transmission, such as airports, schools, hospitals, and crowded places. However, some individuals continue to neglect wearing masks in these locations, leaving the area vulnerable to disease spread, including COVID-19. Therefore, this study proposes the development of an embedded system based on Raspberry Pi 4 for automatic mask detection using YOLOv4 deep learning and a PyQt5 interface. The system is designed to be simple and compact, featuring a user-friendly GUI to effectively detect mask usage on multiple faces in a single detection. Experimental results on 40 samples captured in real-time, with 4 samples taken per image capture and various mask colors and three mask-wearing angles, demonstrated that the average precision, recall, and F1_score rates were each 100%. This outcome proves that the proposed embedded system successfully detects masks on multiple faces with different colors and angles in a single detection with excellent accuracy.