Face Mask Detection using YOLOv8

During the COVID-19 pandemic, ensuring public health and enforcing mask mandates were crucial to controlling the virus's spread. However, manual monitoring methods were not efficient enough. Automated methods that track mask compliance offer a more effective and precise solution for ensuring the safety of the public. This study employs the YOLOv8 algorithm to detect the presence of face masks utilizing two datasets: the AIZOO face mask dataset and the FMD dataset. Three YOLOv8 models (n, s, m) were applied to the AIZOO dataset. YOLOv8m achieved the highest achieved mAP 50 of 95.56%, outperforming other state-of-the-art models in the conducted comparison. The second experiment used the FMD dataset with fine-tuned pre-trained YOLOv8m weights, achieving mAP 50 of 88.98%. The results show that YOLOv8 performs impressively well in face mask detection, proving its reliability and effectiveness across different scenarios.