My main research interest focus on the design and prototyping of diagnostic and prognostic tools based on multimodality or multiparametric 3D images mainly for brain and cancer imaging.
Diagnostic and prognostic tools are aimed at assisting radiologists facing the continuously increasing amount of imaging data that are daily generated in hospitals. While these models target different objective
- Prognosis models aiming at predicting the course and outcome of disease processes,
- Diagnosis models aiming at detecting, localizing and characterizing pathological patterns in the data-,
they both rely on similar theoretical foundations.
This methodological research on CAD systems drives upstream research in the field of machine learning to propose novel algorithms that account for the specificity of medical imaging data including, for instance, uncertainty about the reference training data, low sample, high dimensional and heterogeneous data. We have been investigating the potential of deep learning models over the past few years.
Please find some recent contributions by clicking on the following topics: