Tech-a-thon is an annual 5-week Hackathon organized by the ECE Society of Birla Institute of Technology, Mesra. We selected the ML/AI domain and worked on the theme- Recognition Systems. We built a Mask Detector which helps in distinguishing among people who are wearing and are not wearing a mask in a crowd using facial recognition system.
The Mask-Detector crops the facial area from the uncontrolled real-time images having differences in face size, orientation and background. This helps in better localizing the person who is violating the facemask norms in public areas. It then recognizes the presence of a mask and displays a message accordingly. If it does not detect a face then it produces a beep sound.
The dataset we used consisited of two categories of images i.e. with and without mask. Our project gives an accuracy of 83.64%.
Wearing a mask is a preventive measure that can be used to control the spread of COVID-19 droplets expelled by an infected individual. The key to control COVID-19 pandemic is to maintain social distancing, improving surveillance and strengthening health systems. Studies reveal that wearing a face mask or other covering over the nose and mouth cuts the risk of Coronavirus spread by more than 90%. The mentioned project efficiently handles the identity detection of faces, violating the mask norms increasing the utility of the system for public benefits.