YOLOv5-Based Mask Wearing Status Detection

Face mask wearing detection technology can effectively supervise and control people wearing masks. The traditional mask detection uses only binary classification labels for training, which has low precision and recall. The mask wearing classification method is redefined, and a mask wearing detection method based on YOLOv5 is proposed. The model algorithm learns finer facial features by processing 24000 collected images with corresponding classification labels. Compared with the traditional two-classification method of mask wearing status, the YOLOv5 mask wearing detection method with multi-classification labels achieves 93.4% detection precision and 97% recall rate, which are improved to 3.1% and 2%, respectively, indicating that the new multi-classification labels of mask wearing can effectively improve the performance of mask detection.