Bart M.N. Smets

Bart M.N. Smets is a University Researcher at Eindhoven University of Technology, affiliated with the Center for Analysis, Scientific Computing and Applications and the Geometric Learning & Differential Geometry group within the School of Mathematics and Computer Science. His research focuses on geometric deep learning, partial differential equations (PDEs), and their applications in image processing and convolutional neural networks (CNNs). Dr. Smets has contributed to advancing PDE-based CNN architectures, data-driven cartan connections for vascular tracking, and geometric analysis frameworks.

He teaches the course Differential Geometry for Image Processing and has presented at international conferences on topics like geometric deep learning and PDE-based morphological convolutions. His work aligns with UN Sustainable Development Goals through innovations in medical imaging and computational methods.

Key research themes include equivariant neural networks, sub-Riemannian geometry applications, and mathematical foundations of deep learning. Over 12 publications since 2019 reflect his interdisciplinary approach combining pure mathematics with practical machine learning challenges.