Social Distancing and Face Mask Detection using Yolov4 and CNN

Abstract

Social distancing and wearing face mask is the urge of year 2022 due the possible spread of coronavirus .
Social distancing is one of the possible approach with which we can continue fighting this pandemic.
Motivated by this notion, the ground plan proposed a deep knowledge based framework for automating
the task of covering social distancing using surveillance tape. We proposed COVID-19 Social Distancing and
face mask detector system which is a one-stage detector, which consists of a machine learning module for
calculating the distance among the different classes ,and a CNN model for detecting whether a person is
wearing mask or not. YOLO stands for You Only Look Once, this algorithm is employed for Object Finding
also as Object Tracking. After we performed recognizing with YOLOv4, we calculated the Euclidean distance
between all the detected boxes and filter out or flag the people that are closest to every other indicating
that they're at risk. The high the Euclidean distance score would be the better model was within the spotting
of objects. The minimum distance to keep while adhering to social distancing is 6 feet , keeping this as the
base for calculating distance, the model was trained and used for object detection as well as object tracking.
For the mask detection a CNN model was made using Deep learning and OpenCV. The custom datasets
were used to understand facemasks and were trained on those datasets for detection and tracking. Predictive
Analysis on the static images as well as in the videos to detect face masks has been done. In addition, the
system also used existing IP cameras(web cameras) combined with computer vision to detect people without
Mask and violence of social distancing.