Face recognition with masks is to judge whether pedestrians in images or video sequences wear masks by computer vision technology. It can be used in some specific occasions that need to wear masks, such as canteens, restaurants and other employees who need to wear masks to work, or pedestrians need to wear masks because of special circumstances.

At present, the mask face recognition system on the market usually detects the face first, then divides the face into regions, statistically analyzes the color information of the region below the face, and then judges whether the face wears a mask.

However, in the actual field application, the inclination angle of human face is different, and the interference of different lights is also different, which leads to the unsatisfactory accuracy of traditional methods. Therefore, the existing technology needs to be improved.