Comparison of Face Coverings Detection Methods using Deep Learning
Muhammad Nizam Abdul Aziz, Sofianita Mutalib, Sharifah Aliman · 2021
End of the year 2019, the world has been shocked by a novel coronavirus disease 2019 (COVID-19), and the world is facing a huge health crisis due to the rapid transmission of COVID-19. World Health Organization (WHO) advised wearing a face cover in public places and crowded areas as a part of a comprehensive element to preventive and control measures to limit the spread of COVID-19. However, it is not easy to monitor people manually in these areas. In this paper, a detector model using deep learning for face coverings detection will be presented based on DEtection TRansformer (DETR) algorithm. The custom dataset has been used through this research, which comprises different face covering such as face mask, face shield, niqab, and purdah. The results concluded that the proposed model achieved the highest accuracy percentage of 92.38% and the highest recall percentage of 86.21% as a detector model. Finally, a comparative result with other face coverings model has been presented at the end of the research. The proposed model has achieved higher accuracy, precision, recall and specificity than the other model.