Mask Detection and Classification in Thermal Face Images

Abstract

Face masks are recommended to reduce the transmission of many viruses, especially SARS-CoV-2. Therefore, the automatic detection of whether there is a mask on the face, what type of mask is worn, and how it is worn is an important research topic. In this work, the use of thermal imaging was considered to analyze the possibility of detecting (localizing) a mask on the face, as well as to check whether it is possible to classify the type of mask on the face. The previously proposed dataset of thermal images was extended and annotated with the description of a type of mask and a location of a mask within a face. Different deep learning models were adapted. The best model for face mask detection turned out to be the Yolov5 model in the ''nano'' version, reaching mAP higher than 97% and precision of about 95%. High accuracy was also obtained for mask type classification. The best results were obtained for the convolutional neural network model built on an autoencoder initially trained in the thermal image reconstruction problem. The pretrained encoder was used to train a classifier which achieved an accuracy of 91%.INDEX TERMS Deep neural networks, epidemic prevention, health infrastructure, mask area detection, mask type classification, thermal imaging.

Robust Face Mask Detection by a Socially Assistive Robot Using Deep Learning

“…In [30], the automated detection of face masks and classification of mask types was presented using thermal images. Two deep learning models were adapted for the detection task: (1) the compact YOLOv5 "nano" model and (2) RetinaNet.…”

“…From the aforementioned literature review, a number of existing mask detection methods have been developed, which have only provided binary classification (mask/no mask) using either 2D RGB images [12][13][14][15][16]18,19,52,55], thermal images [30], video streams [15], or real-time data from webcams [13,16,19,28]. However, there is limited research determining if people are properly covering their noses, mouths, and chins with well-fitted masks [23,31], 55].…”

“…To enforce the mandatory wearing of face masks while reducing the time-consuming and costly task for humans to check people as they enter buildings, outdoor venues, etc., automated monitoring systems have been integrated into security cameras [13], tablets [14], and robots [15][16][17][18][19][20]. Recent detection systems have also used deep learning to classify whether people are wearing or not wearing masks (binary classification) in 2D images and videos [21][22][23][24][25][26][27][28][29][30]. Only a handful of face mask detection systems can detect not only if a 2 of 20 person is wearing a mask, but if they are wearing the mask properly [23,31].…”

YOLOv5-S2C2: An Improved Method of Mask Detection Based on Lightweight

“…However, they did not take into account the degree of lightness of the model. Kowalczyk et al [20] have proposed the research topic of automatic detection of masks to reduce the spread of viruses, and they have used YOLOv5 to classify and localize the mask images for their work.…”