Social Distancing Detection System Using Single Shot Detection (SSD) and Neural Networks
Dipti Jadhav, Gokarna Patil, P. Tekade, Shubhada Tambe · Auerbach Publications eBooks · 2024
Social distancing detection system is a crowd monitoring system that can be used during pandemic situations. Authors propose a social distance detection system where people maintaining a safe distance are framed by the green colour and people who don&s;t are framed by the red colour. In this paper, the Single Shot Detection (SSD) model is used for object detection and object class detection. Single Shot Detection (SSD) uses feature maps that extract the object class from the input image and detect objects using the convolution layers of neural networks. Manhattan distance is used to calculate the distance between the detected objects. If the distance between the detected objects is less than the empirical threshold, the objects are framed by red colour otherwise it is framed by green colour. Authors have implemented this model for real-time as well as for recorded CCTV footage. It can also give the real-time count of social distancing violations.