Applying Machine Learning techniques to support COVID-19 diagnosis

The project is about applying Machine Learning (ML) techniques to help diagnose pneumonia produced by SARS-Cov-2 using chest X-Ray images, especially in countries where the virus is widely spread but have few resources to fight it.

The main objectives of this project are:

  • Create machine learning (ML) models to help diagnose SARS-Cov-2 using chest X-Ray images.
  • Validate the models in clinical studies with real patients in the hospitals in collaboration with the University of Antioquia (UdeA).
  • Create a desktop/web application, based on the aforementioned ML models, to evaluate newly-taken images against the model. This application should be of easy installation in hospital computers, and it would assist the doctors in the diagnosis.

The choice of X-Ray images as input for the ML models is motivated by the fact that X-Ray technology is commonly available, even in small medical centers of areas with limited diagnosis equipment.