A Real-Time System for a Safer Society in the Era of the COVID-19 Pandemic Using New Configurations of YOLO and MobileNet

Abstract

To create a secure environment that supports public safety, the proposed solution called I3S-Covid19 (Intelligence system for a safer society in covid-19) which consists of several parts: 1) extract foreground objects in videos received from surveillance camera. 2) Detect whether a person is wearing a mask or not through the use of data augmentation, transfer learning and new configuration of several models (such as MobileNet and YOLOV3). 3) Calculate the distance between people circulating in public or private places using MobileNet-SSD and YOLOV3 with the Euclidean distance measure. Finally, after evaluating the different solutions in different contexts and on different benchmark datasets, the results obtained represent an empirical validation of the benefit derived from the use of deep learning, the internet of things, and computer vision to minimize the spread of COVID-19.