Research Projects

Showing 18 of 18 projects
01
ActiveGeneral

Development of the MEDomics platform

Open-source platform for integrative multi-omics modeling that supports precision medicine workflows across research teams.

View Project
MEDomics platform
02
ActiveGeneralMEDomics Ecosystem

MEDiml – Medical Image analysis with Machine Learning

Open-source, IBSI-compliant radiomics platform, available as a Python package and a no-code desktop application.

View Project
MEDiml radiomics platform
03
ActiveDoctorateMEDomics Ecosystem

Predictive Performance Precision Analysis in Medicine (MED3pa)

Methodology to identify low-confidence predictions and patient profiles where clinical models are unreliable.

View Project
Method overview
04
ActiveDoctorate

Predictive modeling based on multi-level graphical representations of multimodal healthcare data

Multi-level graph neural networks and reinforcement learning to model multimodal patient data and federated training.

View Project
Project presentation
05
ActiveDoctorate

Multimodal fusion of histopathology and clinical data for survival prediction in oncology

Fusion of whole-slide histopathology images and clinical data to build accurate, well-calibrated and interpretable survival models for cancer patients.

View Project
Multimodal fusion of histopathology and clinical data for survival prediction in oncology
06
ActiveMaster's

Optimizing MEDomics platform: Enhancing Usability for Clinical AI Adoption

Master's project focused on improving MEDomics platform usability, workflows, and clinical validation.

View Project
Project illustration
07
ActiveMaster's

Predicting adverse radiation effects of stereotactic radiosurgery for brain metastases using radiomics and machine learning

Simple radiomic model built from pre-treatment brain MRI and dosimetry maps to identify patients at high risk of adverse radiation effects after stereotactic radiosurgery.

View Project
Illustration of Gamma Knife stereotactic radiosurgery
08
ActiveMaster'sMEDomics Ecosystem

MEDfl: A Collaborative Framework for Federated Learning in Medicine

Open-source federated learning framework that lets hospitals train clinical AI models together without moving patient data, available as a Python package and a no-code desktop application.

View Project
MEDfl federated learning framework
09
CompletedGeneral

Refactoring of the MEDiml Python package

Refactoring of the MEDiml codebase, its docstrings and documentation, which also started the first version of the MEDiml app.

View Project
Schematic diagram of the MEDiml package
10
CompletedGeneral

Development of a methodological review on statistical inference and distributed learning in health

Methodological review detailing statistical inference and distributed learning approaches in health research.

View Project
Distributed learning and inference methods
11
CompletedGeneral

Random forests in a distributed learning context

Bachelor's capstone implementing and evaluating federated random forests in a distributed setting.

View Project
Class diagram
12
CompletedGeneral

Graph neural networks for predicting patient mortality within one year of hospital admission

Graph neural network approach to predict one-year mortality and support end-of-life care discussions.

View Project
Project presentation
13
CompletedMaster's

Development of artificial intelligence techniques for the automated identification of electrical energy assets

AI-assisted detection of electrical fuses and symbols to automate interpretation of engineering drawings and photos.

View Project
Examples of fuse detection
14
CompletedMaster's

Radiomics for the management of small renal masses

Radiomics and clinical data integration to classify benign versus malignant small renal masses.

View Project
Example of a small renal mass
15
CompletedMaster's

Multi-task learning for image classification of renal tumors

Multi-task learning framework for renal tumor malignancy, subtype, and grade classification with MRI data.

View Project
Hard-sharing architecture example
16
CompletedMaster's

Resilient predictive models based on quantitative imaging to guide prostate cancer treatment

Robust imaging-based models for prostate cancer treatment guidance using CT and PET modalities.

View Project
Examples of scans
17
CompletedMaster's

Analyzing the Complexity of Radiomics: Approaches for Simplified Modeling

Systematic study of radiomic feature complexity across imaging modalities for precision oncology tasks.

View Project
Project summary
18
CompletedMaster's

Machine learning strategies for the diagnostic and the prediction of late adverse effects related to childhood lymphoblastic leukaemia treatment.

Machine learning models for late adverse effects in childhood ALL survivors, focusing on VO2max and obesity prediction.

View Project
Project presentation