From Pixels to Insights: Leveraging Class-Dependent Activation Maps for Augmented Medical Image Captioning
Scheltjens, V., Verbeke, W., De Moor, B.
In ReviewBMC Medical Informatics and Decision Making
Vincent Scheltjens
PhD Researcher
STADIUS & LIRIS — KU Leuven
Researching how Federated Learning can enable collaborative AI development in healthcare while enhancing patient privacy, enabling hospitals to train powerful machine learning models without sharing sensitive medical data.
Research
Federated Learning
Privacy-preserving distributed machine learning across hospital networks
Healthcare AI
AI-driven decision support systems for clinical environments
ICU Predictions
Length of stay prediction for critically ill patients
Medical Imaging
Radiology report generation with activation maps
Publications
2026
RAFA: Robust Adaptive Federated Aggregation for 3D MRI Segmentation
Scheltjens, V., Verbeke, W., De Moor, B.
In ReviewBMC Medical Informatics and Decision Making
2025
Understanding Deep Learning Models for Length of Stay Prediction on critically ill patients through Latent space Visualization
Wamba Momo, L.N., Scheltjens, V., Verbeke, W., Rademakers, F., De Moor, B.
JournalComputer Methods and Programs in Biomedicine, 268, Art.No. 108832
DOI: 10.1016/j.cmpb.2025.1088322024
Target informed client recruitment for efficient federated learning in healthcare
Scheltjens, V., Momo, L.N.W., Verbeke, W., De Moor, B.
JournalBMC Medical Informatics and Decision Making, 24(1), Art.No. ARTN 380
DOI: 10.1186/s12911-024-02798-4Augmented Medical Image Captioning: Leveraging Class-Dependent Activation Maps for Enhanced Radiology Report Generation
Scheltjens, V., Wamba Momo, L., Verbeke, W., De Moor, B.
ConferenceIEEE/CVF Conference on Computer Vision and Pattern Recognition 2024 (CVPR), Seattle, United States
2023
Client Recruitment for Federated Learning in ICU Length of Stay Prediction
Scheltjens, V., Wamba Momo, L.N., Verbeke, W., De Moor, B.
ConferenceIEEE International Conference on e-Science, Limassol, Cyprus
DOI: 10.1109/e-Science58273.2023.10254908Client Recruitment for Federated Learning in ICU Length of Stay Prediction
Scheltjens, V., Wamba Momo, L.N., Verbeke, W., De Moor, B.
PreprintarXiv
2022
Multimodal Deep Learning for Early Length of Stay Prediction using Patient Similarity Embeddings
Vandenberghe, A., Wamba Momo, L., Scheltjens, V., De Moor, B.
ConferenceBNAIC/BeNeLearn, Mechelen, Belgium