An Automated Facial Mask Detection System for Minimizing COVID-19 using Multi-Task Cascaded CNN Model


  • Shyam Yadav
  • Vijay Trivadi


An outbreak triggered by a new coronavirus has continued to spread over the globe till today. Nearly every area of development has felt the effects of COVID-19. The state of healthcare is now quite problematic. One of the many preventative steps adopted to limit the spread of this illness is the use of a mask. To limit the spread of COVID-19, we propose a system that identifies persons who are not wearing any face masks in a smart city network that is monitored by CCTV cameras in every public location. It is the cameras’ job to take pictures of people in public locations, and the photos they acquire are sent into a system that looks for anyone without a face mask. Anyone found without a face mask will have their identity reported to the appropriate authorities, who will take the appropriate measures. Authorities are alerted via the city’s network whenever an unmasked individual has been found in the area. On a face mask dataset that includes photos of individuals under different masks, deep learning-based MTCNN architecture is implimented. People with and without face masks were identified with 99.76% accuracy by the trained architecture in previously unreported test data. It is believed that the findings of this research would help many nations throughout the globe combat the spread of this contagious illness.