Geometric and variational methods for diffusion tensor MRI processing

Christophe Lenglet 1
1 ODYSSEE - Computer and biological vision
DI-ENS - Département d'informatique de l'École normale supérieure, CRISAM - Inria Sophia Antipolis - Méditerranée , ENS Paris - École normale supérieure - Paris, Inria Paris-Rocquencourt, ENPC - École des Ponts ParisTech
Abstract : This thesis deals with the development of new processing tools for Diffusion Tensor Magnetic Resonance Imaging (DT-MRI). This recent MRI technique is of utmost importance to acquire a better understanding of the brain mechanisms and to improve the diagnosis of neurological disorders. We introduce new algorithms relying on Riemannian geometry, partial differential equations and front propagation techniques. The first part of this work is theoretical. After a few reminders about the human nervous system, MRI and differential geometry, we study the space of multivariate normal distributions. The introduction of a Riemannian structure on that space allows us to define statistics and intrinsic numerical schemes that will constitute the core of the algorithms proposed in the second part. The properties of that space are important for DT-MRI since diffusion tensors are the covariance matrices of normal laws modeling the diffusion of water molecules at each voxel of the acquired volume. The second part of this thesis is methodological. We start with the introduction of original approaches for the estimation and regularization of DT-MRI. We then show how to evaluate the degree of connectivity between cortical areas. Next, we introduce a statistical surface evolution framework for the segmentation of those images. Finally, we propose a non-rigid registration method. The last part of this thesis is dedicated to the application of our tools to two important neuroscience problems: the analysis of the connections between the cerebral cortex and the basal ganglia, implicated in motor tasks, and the study of the anatomo-functional network of the human visual cortex.
Document type :
Complete list of metadatas
Contributor : Estelle Nivault <>
Submitted on : Wednesday, February 17, 2010 - 1:54:18 PM
Last modification on : Friday, May 25, 2018 - 12:02:04 PM
Long-term archiving on: Friday, June 18, 2010 - 9:11:29 PM


Files produced by the author(s)


  • HAL Id : tel-00457463, version 1



Christophe Lenglet. Geometric and variational methods for diffusion tensor MRI processing. Human-Computer Interaction [cs.HC]. Université de Nice Sophia Antipolis, 2006. English. ⟨tel-00457463⟩



Record views


Files downloads