Felix Ambellan

Felix Ambellan is a researcher at the Zuse Institute Berlin (ZIB), working in the Visual and Data-Centric Computing department within the Mathematics of Complex Systems division. His research focuses on applying advanced mathematical and computational techniques to medical imaging problems, particularly in the context of knee osteoarthritis and neurological disorders.

Dr. Ambellan completed his doctoral studies at Freie Universität Berlin, where he earned his PhD in 2022 with a thesis titled "Efficient Riemannian Statistical Shape Analysis with Applications in Disease Assessment," supervised by Christof Schütte and Christoph von Tycowicz.

His primary research interests lie at the intersection of medical imaging, computational geometry, and machine learning. Ambellan specializes in statistical shape analysis using Riemannian geometry, developing novel approaches for disease assessment through anatomical shape variations. His work has significant applications in knee osteoarthritis diagnosis and Alzheimer's disease grading, where he applies graph neural networks and manifold-valued statistics to extract clinically relevant information from medical images. He has made substantial contributions to the field of statistical shape modeling, particularly through the development of the open-source Python library Morphomatics.

Analysis of Ambellan's recent publications reveals a strong focus on advancing statistical shape modeling techniques within non-Euclidean spaces. His work bridges theoretical mathematics with practical medical applications, particularly in developing methods that can handle the complex geometry of anatomical structures. Many of his papers demonstrate how incorporating geometric awareness into machine learning models improves diagnostic accuracy for conditions like knee osteoarthritis. His research often involves large-scale medical image datasets, including thousands of knee MRI scans from the Osteoarthritis Initiative.

Dr. Ambellan is actively involved in several research projects including "Manifold-Valued Graph Neural Networks," "Morphological Scoring of Disease States," and "Treating Osteoarthritis in Knee with Mimicked Interpositional Spacer," which reflect his commitment to translating theoretical advances into clinical applications. His work has been published in high-impact journals including Medical Image Analysis, Physics in Medicine and Biology, and BMC Medical Imaging, demonstrating both the theoretical rigor and practical relevance of his research.