Real-Time AI-Based Face-Mask Detection

The 2019 coronavirus (COVID-19) has recently caused a serious worldwide health disaster due to its rapid proliferation. To address the problem of spreading COVID-19, the World Health Organization (WHO) published a few strategies. A COVID-19 preventative strategy is to cover your face with a mask or a piece of cloth in public and busy settings. Manually observing and segregating people wearing masks from those who are not wearing masks is really difficult. Therefore, the primary goal of this research study is to construct a face mask detection model as an embedded model by using two architectures, viz., Mobile Net and YOLO. When someone ignored social norms or did not wear masks, it was revealed. As a final step, the authors deployed the hybrid model on a Raspberry Pi to act as a prototype that will detect whether an individual has put on a mask or not. The proposed model achieved a 99% confidence rate. The comparison of multiple face detection and face mask classification methods is another feature of this work. The accuracy of the system is 99%. As a result, our approach really records people wearing or not wearing masks. As a result, our technology detects whether an individual wore a mask or not in a real-time situation and permits access only to the individual who wore a mask. As a future scope, the current Raspberry Pi module along with the Microsoft camera can be deployed in a range of retail establishments, an office complex, or at airport gates.