The Novel Coronavirus outbreak worldwide has made people to consider different ways to perhaps protect themselves from infections. Several safety precautions, including avoiding crowded areas, maintaining hand hygiene, and avoiding touching eyes, mouth, or nose. Researchers, mathematicians, pharmacists, and other professionals have all been challenged by the pandemic to find solutions to the pandemic situation. Algorithms and concepts related to machine learning also find a great place for many scientists. Among all the preventive measures, social distancing is one of the most important protective methods for flattening the COVID-19. An efficient and cogent social distancing violation analyzer tool would play a crucial role in detecting humans if they are not maintaining social distance. The system examines frames it receives as input to find humans in the visual frame. The pairwise distance between each recognized individual is then calculated, and based on the distance value acquired, the system will issue an alert for those who are not maintaining the social distance. To determine how well the system performs, it is tested using a variety of image input acquisition techniques, including camera, video, and image. The proposed system achieves a greater precision for videos, according to experimental findings. Applications like human tracking, pedestrian recognition, and vehicle tracking can all be added to the system.

Keywords : Social Distancing, Pedestrian Detection, Convolutional Neural Network, YOLO, COVID-19, Object Detection