ERFM: An Efficient and Robust Detection Algorithm for Face Mask

In order to prevent and respond to sudden public health emergencies such as COVID-19, it has become increasingly important to accurately detect in real time whether people in public places are wearing masks. Previous deep learning-based mask detection algorithms suffer from low accuracy, poor real-time performance and robustness. To solve these problems, in this paper, we propose an efficient and robust detection algorithm (ERFM) for real-time face mask detection based on You Only Look Once version 8 (YOLOv8). Our ERFM algorithm consists of an efficient face mask detection algorithm (EFM) and a robust face mask detection algorithm (RFM). ERM, which is based on YOLOv8 object detection algorithm, is used for high-precision and real-time face mask detection task, while EFM, which is based on YOLOv8 pose estimation algorithm, is used to defend against mask imitation attacks. In comparison with other mainstream face mask detection algorithms, our ERFM shows significant advantages in robust, accuracy and speed.