Face Recognition of People Without Masks to Control COVID-۱۹ Using Deep Learning

The world is still facing a major health crisis due to the rapid transmission of COVID-۱۹ (coronavirus) and its vast, varied, and different strains. The World Health Organization has issued several guidelines to protect against the spread of Covid-۱۹. According to the World Health Organization, a preventive measure to prevent COVID-۱۹ is to use a mask in public and crowded places. Manual monitoring is complicated in these areas. In this study, a transfer learning model is used to automate the process of detecting people who do not have a mask. The proposed model is built by fine-tuning the advanced deep learning model of MobileNet pre-trained. The proposed model is trained and tested on Face Mask ۱۲k Images Dataset. For better training and testing, data preprocessing techniques have been applied to the target data set and applied to the sorted data for validation. According to the evaluation of the proposed results and the improvement of the accuracy criterion up to ۹۸%, it can be said that the proposed method has performed better than the methods used in recent studies.

کلیدواژه ها:

Deep learning ، transfer learning ، Covid-۱۹ ، facial recognition ، MobileNet