This is the authors' abstract. We don't add key points for this paper.
To restrict COVID-19, individuals must remain two meters away from one another in public since public health authorities find this a healthy distance. In this way, the incidence of “social distancing” keeps pace with COVID-19 spread. For this purpose, the proposed solution consists of the development of a tool based on AI technologies which takes as input videos (in real time) from streets and public spaces and gives as output the places where social distancing is not respected. Detected persons who are not respecting social distancing are surrounded with red rectangles and those who respect social distancing with green rectangles. The solution has been tested for the case of videos from the two Holy Mosques in Saudi Arabia: Makkah and Madinah. As a novel contribution compared to existent approaches in the literature, the solution allows the detection of the age, class, and sex of persons not respecting social distancing. Person detection is performed using the Faster RCNN with ResNet-50 as it is the backbone network that is pre-trained with the open source COCO dataset. The obtained results are satisfactory and may be improved by considering more sophisticated cameras, material, and techniques.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Alyami et al. (2021) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: