Enhanced Deep Convolutional Neural Networks for Facial Mask Recognition using MobileNetV2 Transfer Learning Techniques

COVID-19., the corona virus which is an infectious disease that affect the day-to-day life of every individual person in the world. Masks must now be worn in public spaces to reduce the spread of the disease. So., detecting face mask is very important in public places., crowded areas like shopping malls., workplaces., public transport. The existing models may not be able to predict the output for the images with reflective background., require heavy computational load and provide less accuracy. Face mask identification is accomplished in using machine learning techniques such as Tensorflow., Keras., and OpenCV. This proposed model recognizes faces in images or videos and makes predictions about whether a person or group of individuals is wearing a mask correctly., wrongly., or not at all. It also distinguishes between faces with just masks and faces that are covered by additional things. Convolutional Neural Networks (CNN) are used in the proposed model to determine the output and train the dataset. The system's accuracy was 95%.