Computer vision technologies for social distancing

Many technological applications have been developed and implemented in the last two years to fight the COVID-19 pandemic via social distancing. Despite its importance in response to the coronavirus, the remaining challenge is the limitation of human resources to monitor and provide timely warnings to maintain appropriate activities, such as keeping distance between each other, wearing face masks, or complying with curfew restrictions. To tackle this problem, computer vision researchers have proposed numerous approaches for autonomous object detection and distance measurement, which will be summarised in this chapter. First, vision-based applications in intelligent surveillance systems for social distancing monitoring as well as masked face detection, are introduced. Then, core classical and modern neural network-based methodologies for these applications are analysed. A simple masked face detection is developed to verify its effectiveness and limitations, followed up by remarks and discussions on open problems.