Virtual-reality-based digital twin of office spaces with social distance measurement feature

Social distancing is an effective way to reduce the spread of SARS-CoV-2 virus. Many students and researchers have already attempted use computer vision technology automatically detect human beings in field view a camera help enforce social distancing. However, because present lockdown measures several countries, validation systems using large-scale datasets challenge. In this paper, new method proposed for generating customized validating deep-learning-based models virtual reality (VR) technology. Using VR, we modeled digital twin (DT) existing office space used it create dataset individuals different postures, dresses, locations. To test solution, implemented convolutional neural network (CNN) model detecting people limited-sized real humans simulated humanoid figures. We detected number persons both synthetic with more than 90% accuracy, actual measured distances were significantly correlated (r=0.99). Finally, intermittent-layer- heatmap-based data visualization techniques explain failure modes CNN. A application DTs enhance workplace safety by measuring distance between individuals. The our pipeline along DT shared visualizing environmental behavior aspects preserves privacy improves latency such monitoring only extracted information streamed.