Variational Approaches to the Estimation, Regularization and Segmentation of Diffusion Tensor Images

Rachid Deriche (1) , David Tschumperlé (2) , Christophe Lenglet (1, 3) , Mikaël Rousson (1)

1 ODYSSEE - Computer and biological vision
2 Equipe Image - Laboratoire GREYC - UMR6072
3 Siemens Corporate Research

Diffusion magnetic resonance imaging probes and quantifies the anisotropic diffusion of water molecules in biological tissues, make it possible to non-invasively infer the architecture of the underlying structures. In this chapter, we present a set of new techniques for the robust estimation and regularization of diffusion tensor images (DTI) as well as a novel statistical framework for the segmentation of cerebral white matter structures from this type of dataset. Numerical experiments conducted on real diffusion weighted MRI illustrate the technique and exhibit promising results.

Domaines: Traitement des images [eess.IV]

Rachid Deriche, David Tschumperlé, Christophe Lenglet, Mikaël Rousson. Variational Approaches to the Estimation, Regularization and Segmentation of Diffusion Tensor Images. Paragios, Chen & Faugeras. Mathematical Models of Computer Vision: The Handbook, Springer, pp.517--530, 2005. ⟨hal-00336537⟩