Jesús Tomás and Jaime Lloret, experts in mobile application development and artificial intelligence at Campus Gandia of the Universitat Politècnica de València (UPV) are in charge of a research project for the development of a mobile application to detect the proper placement of masks using artificial intelligence. This project is being carried out in collaboration with the surgical nursing group of the Hospital General d’Ontinyent (GEQUO), a FISABIO affiliate. The joint collaboration in this project is rooted on three research fields: artificial intelligence, health prevention and app development.
This application rises from the need to increase safety in the use of the mask and to raise awareness on the importance of its proper placement. Masks are the most effective complement to the recommended physical distance and hygiene measures to reduce the risk of transmission and infection of COVID-19. The effectiveness of the masks depends on the follow-up of some general hygiene measures, among which the adequate coverage of the mouth and nose stands out, blocking the droplets of large particles and splashes that may contain germs.
With the INTEMASC app, the user will be able to verify the fit of the mask through the mobile camera quickly and easily. To this end, the research team is studying the implementation of artificial intelligence techniques to indicate to the user the possible problems with mask placement.
The application will detect hygienic and surgical masks, although this will depend on the results obtained in the experimentation phase, and can be extended to other types.
The final objective of the project is to develop an image recognition system that will detect if a mask is fitted properly. The objective of the first stage is to create a corpus that trains neural networks through an Android application that will streamline this process.
In the development of machine learning models and, especially, of artificial neural networks, it is not necessary to know the precise solution to a problem, but it does require a considerable number of examples indicating how it has been solved. Therefore, the photographs taken by the users through the app will be collected with the mask worn properly and improperly. After collecting the photographs, a group of experts will label them indicating the placement problems detected by a second mobile application.
The problems to be detected are: no mask, non-symmetrical, loose, under the bridge of the nose, under the nose, underneath glasses, neck adjustment, beard, improper nose placement, improper elastic band placement, side gap.
Eventually, the final version of the beta application for Android will be obtained, which will be implemented for iOS after testing and validating.
COLLABORATE WITH THE PROJECT
You can collaborate with the project by downloading the free INTEMASC application in the Play Store (only for Android) and taking selfies wearing the mask properly and improperly. A database will be created with all the photos, which used to develop the final version of the image recognition application to verify the fit of your mask.
PROTECCIÓN DE DATOS: Las fotografías son tratadas de forma totalmente anónima y solo serán utilizadas para los fines descritos en el proyecto. El uso de la aplicación supone la aceptación de las políticas de privacidad por parte del usuario.
DATA PROTECTION: The photographs will treated completely anonymously and will only be used for the purposes described in the project. The use of the application implies the acceptance of the privacy policies on the part of the user.
In addition to raising awareness of the importance of proper mask placement, the project aims to achieve research results in artificial intelligence and to obtain statistical data on mask use, which will enable the development of future innovative solutions in these fields of research.
RDI Management and Promotion Officer at Campus Gandia