SDSC Develops a Deep Learning and Computer Vision-based Method to Monitor Social Distancing in Real Time

Movie 1: Social distance monitoring in a public area. Social distances less than 2 meters are highlighted.

Movie 2: Social distance monitoring on a basketball court. Social distances less than 2 meters are highlighted.

The COVID-19 pandemic has brought unprecedented challenges for Singapore and the world. Practicing social distancing is considered one of the most effective strategies to contain coronavirus. After easing the restrictions on social activities in Phase II, an increasing number of gatherings of large crowds were spotted and reported in different places, such as sport courts, void decks, and public areas in front of bars and restaurants.

In this project, we have developed a deep learning and computer vision-based strategy to measure the distances between people in real time. This approach can be integrated with existing CCTV systems to monitor social distancing and highlight violations of safety rules.

If there are any interests in using or adapting the developed tools in this project, please contact SDSC at sdsc@nus.edu.sg.

Project Title: Social Distance Monitoring in Real Time
Done By: Singapore Data Science Consortium (SDSC)