Computer Vision Based Face Mask Recognition in Religious Mass Gatherings and COVID-19 Infection

Religious mass gathering at religious places presents a great difficulty in crowd management, providing the necessities of basic amenities, and addressing their healthcare requirements. Certain ritual practices may increase the risk of transferring respiratory pathogens. The COVID19 pandemic was an unexpected healthcare crisis with impacts reflected in every sphere of human lives. To limit the spread of the virus, the World Health Organization (WHO) declared wearing face masks is imperative manner. But manual assessment, whether an individual wears face masks or not in a public place, will be a difficult task. The need to monitor people wearing face masks is required to construct an automatic method. This study introduces a Computer Vision based Face Mask Recognition in Religious Mass Gatherings and COVID-19 Infection, named the CVFMR-RMG technique. The presented CVFMR-RMG technique aims to recognize the faces with and without masks via CV and image processing techniques. Initially, the input images are pre-processed via the adaptive Weiner filtering (WF) model. In the presented CVFMR-RMG technique, the face detection process takes place using Faster Region Convolutional Neural Network (Faster RCNN) approach. For facemask detection and classification, the sparse stacked auto encoder (SSAE) technique was employed. To improve the facemask detection performance of the SSAE model, a modified artificial fish swarm algorithm (MAFSA) is used for hyperparameter tuning. The performance assessment of the CVFMR-RMG system can be tested using the facemask database. The simulation outcomes illustrate the improved performance of the CVFMR-RMG system over other models.