Face Mask Detection Based Entry Control Using XAI and IoT

Abstract

Today it has become mandatory for all the citizens to wear a face mask to protect them from COVID-19. Also taking two doses of vaccine is a must to visiting public places and currently, the only method to verify whether a person is fully vaccinated is by showing a vaccine certificate. The proposed application is helpful for elderly people who find it difficult to use smart phones. The shop owners, offices, banks, or any public place can check for restrictions of entry if anyone is not wearing a mask. As a result, no need for any guard to keep an eye on people. Machine learning techniques with Explainable AI (XAI) can solve these problems easily and results are made understandable to end-users because of the explaining ability and interpretability of neural network models. The system performs well for prediction and gives more accurate and trustworthy predictions. Hence XAI is more reliable in healthcare systems. The proposed system is implemented completely on Raspberry Pi allowing a complete embedded application. The application is developed using Python and HTML. PyCharm/Visual Studio Code with the help of an open-source library is used for training, defining, etc. Machine learning models used for the system are Tensorflow.js, Keras, OpenCV, etc. The whole application can run on a microcontroller such as Raspberry Pi, which allows one to simply plug and play the system at any time.