Deep Learning Algorithms to Diagnose COVID-19 on Chest X-rays

The objective of this project is to develop an artificial intelligence tool based on deep learning algorithms that allows identifying, by means of chest radiography, the presence of lung involvement, even in incipient stages, caused by COVID-19. This would allow an automated detection system for COVID-19 in suspected patients. It will be carried out using a database of 740 radiographs of patients from the San Cecilio de Granada University Hospital, half healthy and the other half with COVID, and will be expanded with data from other national and international hospitals. In addition, the tool will be able to distinguish between COVID and other lung diseases, such as bacterial pneumonias, other viral pneumonias, tumors, etc. In this way, Any health center that has X-rays will be able to have a COVID probability alert and start the protocols in advance. 56 members of four Andalusian and Galician hospital centers will participate in the work, and three other hospitals in the national territory and a university center will participate as collaborating entities.