Person-to-person transmission plays a key role in the spread of COVID-19 and its variants, as it spreads through small droplets that are expelled when a pre-symptomatic patient sneezes, coughs or simply speaks. Therefore, in order to reduce the rates of spread of the virus, the use of face masks in public areas is essential.

In general, monitoring people who do not respect the use of face masks is a challenging task, especially in places with large crowds where it is impossible to track if everyone is wearing a mask. For this reason, there is a great need for a computer system that allows for the automatic tracking of the correct/incorrect use of face masks.

For such purpose, students Fabricio Crespo and Brian Crespo presented the project “A Computer Vision Model to Identify the Incorrect Use of Face Masks for COVID-19 Awareness”, under the direction of professor Eugenio Morocho, PhD., all of them members of the “DeepARC” research group of the School of Mathematical and Computational Sciences. The project, which was published in the scientific journal Applied Sciences, was carried out in collaboration with the Muhammed VI University in Morocco and the University of Cauca in Colombia.

The project proposes the creation of a composite convolutional neural network (CNN) architecture based on two computer vision tasks: object localization to discover faces in images/videos and CNN image classification to categorize faces and show if someone is wearing a mask correctly, incorrectly or not wearing one at all.

The first CNN is based on RetinaFace, a model for detecting faces in images, whereas the second CNN uses a ResNet-18 architecture as a classification backbone, allowing accurate identification of people who are not correctly following the COVID-19 health recommendations on the use of masks.

In order to give this technology a better and greater global use, the “DeepARC” research group made available to the public both the data set used to train the classification model and the artificial vision code proposed and optimized for the implementation of integrated systems in any part of the world.

With these actions, Yachay Tech contributes to the development of knowledge based on technology and innovation and contributes to the community.