MED-AI

In the frames of the AI4EU (https://www.ai4europe.eu/) project the MED-AI solution is an early prototype which might be substantially extended in terms of functionality towards a full-fledged platform and a Clinical Decision Support System (CDSS) according to the E-AI Challenge of Centre Hospitalier Universitaire de Liege (CHUL) and the users’ needs. With the grant the MED-AI solution (with restricted functionality) was developed, tested and validated (supported by AI4EU and CHUL) to reach TRL7 in an operational environment. The project aims at providing an AI and data driven solution which might pave the way of the CHUL’s transition towards personalised medicine (PM). Main objective is to build a software platform for early detection and diagnosis of Covid-19 and pneumonia. The platform is Web based, built according to principles of adaptive design and can be accesses from PCs and mobile devices. Main input data types are Computer Tomographic (CT) or X-ray images. The solution uses deep convolutional neural networks. Two approaches were explored: training custom neural network and transfer learning.