Structure and Attention Enhancement of YOLO based Mask Wearing Recognition

The global COVID-19 pandemic dramatically changes our mod-ern daily life, e.g., wearing a facial mask. Thus, mask wearing recognition brings a new challenge of deep learning. This paper proposes structure and attention enhancement of YOLO which is one famous real-time object detection algorithm. Based on our initial experimental results, the proposed enhancement can in-crease 9.1% accuracy and reduce 15.8% computational effort.