Wed November 6th 2019
16:00 – 17:00
Seminar Biophysical and Data Driven AI for Medical Image Analysis
Hervé Delingette


There is a strong need to assist physicians for prevention, diagnosis, prognosis and therapy in medicine through the use of computational tools. Machine learning and especially supervised learning plays a growing role in the design of those tools but is also limited by a potential lack of explainability, and generality. In this lecture, I will present some issues related to pure data-driven AI methods when performing image segmentation, registration and synthesis and ways to partially overcome them. I will also show how models based on the law of physics to describe image formation, deformation of soft tissues or heat transfer can improve the augmentation of databases of images for supervised learning. Conversely, machine learning can also contribute to the assimilation of data into a personalized biophysical models."

Biography: Hervé Delingette is a Research Director at Inria, director of an academy of excellence of Université Côte d’Azur, a member of the Board of the MICCAI Society, and he is holding a Chair at the new AI institute 3IA Côte d’Azur. He received his engineering degree from Ecole Centrale Paris and his Phd from Inria. His research focuses on various aspects of artificial intelligence in medical image analysis, computational physiology, and surgery simulation.
Go back to the agenda.

The 10th Complex Motion in Fluids 2021
Max Planck Gesellschaft
Centre for Scientific Computing