Single shot multibox detector for detecting and classifying masked faces in digital images

This paper presents a method for detecting and classifying masked faces in digital images using the Single Shot Multibox Detector (SSD). The location of faces in an image and their category, indicating whether the detected face is wearing a medical mask, wearing it improperly, or not wearing one, are found based on the proposed method. This detection is used for security to find the location of faces in a certain category. SSD is a single-stage detector that utilizes various boxes of different sizes in its architecture to detect the location of the target in the image and also the category of the object within the boxes. The various boxes are used to detect different face sizes in the image. The experiments are conducted on three datasets, they are AFLW (Annotated Facial Landmarks in the Wild) with 1865 images, MAFA (Masked Face Analysis) with 3616 images, and MakeML with 853 images. The mAP (Mean Average Precision) of the method in the detection is 41.52%.