Maintaining the use of face maks in an indoor establishment conforming with the COVID-19 protocols thru computer vision model: an approach to detect mouth as region of interest

Abstract

Effective COVID-19 pandemic control strategies need ongoing attention to prevent negative effects on public health and the global economy. WHO suggests several strategies to reduce infection rates and stop the depletion of limited medical resources in the absence of effective antivirals. Wearing a mask is one of the non-pharmaceutical therapies that can be used to reduce the main source of SARS-CoV2 droplets released by an infected person. All governments require that people wear masks that cover their mouths and noses, regardless of differences in medical resources and mask styles. The suggested mask detection model might be a useful tool for ensuring that safety precautions are adhered to in the upcoming years. This research study develops a mask detection model using the deep learning model and the most recent object identification technique. The goal of the study is to ascertain if detecting mouths rather than facemasks is a better substitute for identifying the incorrect wearing of face masks, as the mouth is always visible when a face mask is worn improperly. The excellent performance of the proposed model makes it perfect for video surveillance equipment. The recommended method focuses on using data augmentation methods to transform a dataset of 500 images into an enriched dataset. The mean average accuracy of the Data Augmentation-based Mask Detection Model was determined to be 92.4%.