Hierarchical Segmentation of Multiple Sclerosis Lesions in Multi-Sequence MRI

Automatic segmentation of multiple sclerosis lesions in magnetic resonance images remains a challenging task. In this study, we present a fully automatic method to extract lesions from multi-sequence MRI (T1, T2, T2 FLAIR, Proton Density) within an EM based probabilistic framework. The method uses the available MRI sequences in a hierarchical, orderly manner. First the T2 FLAIR sequence is used to generate a segmentation of supra-tentorial lesions. Then T2 and T1 lesion loads are computed, providing an insight into lesion structure. A priori anatomical knowledge is incorporated in the form of a probabilistic brain atlas.

• Type: Conference
• Year: 2004
• Where: ISBI
• Authors: Guillaume Dugas-Phocion, Miguel Ángel González Ballester, Grégoire Malandain, Nicholas Ayache, Christine Lebrun, Stéphane Chanalet, Caroline Bensa
• Keywords: ISBI 2004, Medical Imaging, Multiple Sclerosis Lesions, Priori Anatomical Knowledge, Probabilistic Brain Atlas