Social Distancing and Mask Detection Tool using Deep Learning .

Covid-19 cases witnessed worldwide till date is 132M and the recovered cases are 74.7M and death count as of 06/04/2021 is 2.86M. By this statistics it is evident that the need of the hour is to control virus spread somehow. It is still not very safe even after the vaccination is found. The only option left is to stop further spread of virus and in order to control spread one has to maintain minimum social distance and has to wear face mask when in public places. The proposed model includes real time video input which is first applied with the YOLOv3 algorithm to get the bounding boxes for a particular person. The output of YOLOv3 is further considered to obtain the positions of people and also to detect faces using Dual Shot Face Detector (DSFD). The positions calculated are subjected to DBSCAN to detect clusters and the faces detected are fed to a face mask classifier. Finally, the status of face mask detection and social distancing is displayed with the respective bounding boxes.