A Computer Vision System to Help Prevent the Transmission of COVID-19

Eyiokur, Fevziye Irem; Ekenel, Hazim Kemal; Waibel, Alexander

The COVID-19 pandemic affects every area of daily life globally. To avoid the spread of coronavirus and retrieve the daily normal worldwide, health organizations advise social distancing, wearing face mask, and avoiding touching face. Based on these recommended protective measures, we developed a deep learning-based computer vision system to help prevent the transmission of COVID-19. Specifically, the developed system performs face mask detection, face-hand interaction detection, and measures social distance. For these purposes, we collected and annotated images that represent face mask usage and face-hand interaction in the real world. We introduce two different face datasets, namely, Interactive Systems Labs Unconstrained Face Mask Dataset (ISL-UFMD) and Interactive Systems Labs Unconstrained Face Hand Interaction Dataset (ISL-UFHD). We trained the proposed models on our own datasets and evaluated them on both our datasets and already existing datasets in the literature without performing any adaptation on these target datasets. Besides, we proposed a distance measurement module to track social distance between people. Experimental results indicate that ISL-UFMD and ISL-UFHD represent the real-world's diversity well. The proposed system achieved very high performance and generalization capacity in a real-world scenario for unseen data from outside the training data to detect face mask usage, face-hand interaction detection, and measuring social distance. The ISL-UFMD and ISL-UFHD datasets will be available at https://github.com/iremeyiokur/COVID-19-Preventions-Control-System.