Performance Evaluation of Face Mask Detection for Real-Time Implementation on an Rpi

Mask-wearing remains to be one of the primary protective measures against COVID-19. To address difficulty manual compliance monitoring, face mask detection models considerate both frontal and angled faces were developed. This study aimed test performance said in classifying multi-face images upon running on a Raspberry Pi device. The accuracies inference speeds measured compared when inferencing with one, two, three desktop Pi. With an increasing number image, models’ observed decline, while their not significantly affected. Moreover, YOLOv5 Small model was regarded potentially best for use lower resource platforms, as it experienced 3.33% increase accuracy recorded least time two seconds per image among models.