A Solution to COVID-19: Detection and Recognition of Faces with Mask

Nithyashree, V.; Roopashree, S.; Duvvuri, A.; Vanishree, L.; Madival, D. A.; Vidyashree, G.

In this COVID-19 crisis, wearing masks is necessary and no longer an option to the general public. To follow the strict directives given by the government, the businesses have to implement a cost-effective approach to ensure that all its employers wear a face mask and help to control the spread of coronavirus. The proposed solution consists of an automatic face mask detection system that eliminates the need of an employee at the entrance. The working model detects a face mask in every person by analysing each frame of the video and alert through security mail when the mask is not detected. Our proposed system is designed using the Convolution Neural Network (CNN) model for mask detection and image subtraction technique to recognise the faces with mask. The scope of the project pertains to avoid the entry of unauthorized people into an organization by recognizing the face despite the presence of a mask. The model shows an accuracy of 99.82% on a custom dataset. It is an effective protection step to impede the transmission of the novel coronavirus. © 2021 IEEE.