An automated face mask detection system using transfer learning based neural network to preventing viral infection

As the “Internet of Medical Things (IoMT)” grows, healthcare systems can collect and process data. It is also challenging to study public health prevention requirements. Virus transmission can be prevented by wearing a mask. The World Health Organization (WHO) recommends wearing a facemask to protect against the COVID-19 pandemic—the levels of a pandemic rise across almost all regions of the world. By following the WHO rules, we support the development of face mask-detecting technologies and determine whether or not people are using masks in public locations. The proposed paradigm in this paper will work in three stages. Firstly, we use an Image data generator to import the images. In addition to using a Haar cascade (HC) classifier for detecting faces, residual learning (ResNet152V2) trains a model that detects whether someone is wearing a face mask. Detection and classification are carried out in real-time with high precision. Compared with other recently proposed methods, the model achieved 99.65% accuracy during training and 99.63% during validation.