Vision transformer-based deep learning techniques for improper face mask-wearing detection

COVID-19 pandemic causes a global catastrophe that remarkably affects individual lives and society, as well as the economy. The world has taken numerous defenses against these contagious diseases and the use of facemasks is one of the most crucial defense mechanisms. However, this preventative strategy is fruitful if the face masks are worn correctly. Unfortunately, only less than 25% of people use masks correctly. In this paper, we proposed to apply a most recent deep learning-based image classification model named vision transformer to automatically detect improper face mask-wearing, i.e the facemasks are being worn correctly or not. In contrast with the traditional image processing or deep learning model, the vision transformer base model shows a significant impact on incorrect face mask detection. We did the experiment on a large custom dataset consisting of 2,03,780 digital images with 3 class labels namely With Mask (when people are wearing the mask properly), Without Mask (when people are not wearing the mask) and Incorrect Mask (when people are not wearing the mask properly) to train, validate and test vision transformer model that can classify the use of masks correctly. The results show that the accuracy achieved with pre-trained vision transformer models is highly remarkable.