<![CDATA[During the Covid-19 pandemic, students were required to wear masks in classroom learning. However, students often do not use masks, so they are prone to transmission of Covid-19. For this reason, this study proposes the development of a real-time multi-face mask automatic detection system in classroom learning using YOLOv4 deep learning. Experimental results on 22 samples of students who collected real-time/live video data every 3 minutes for 20 scenarios proved that the proposed system was successful in detecting objects wearing masks (PM) and not wearing masks (TPM) with the average percentage of precision was 95.63% for PM and 97.33% for TPM, the average percentage of recall was 61.61% for PM and 60.23% for TPM, and the average percentage of F-measure was 74.55% for PM and 74.00% for TPM. This results indicate an effective, valid and accurate proposed system for monitoring the use of masks in classroom learning easily and automatically.]]>
Copyrights © 2024
