"White matter bundle analytics: building atlases and tractometry in the space of streamlines"

15th of December, 2015, 11:00-12:00, CONFERENCE ROOM OF CGFB

A conference by Pr Maxime Descoteaux, Associate Professor, Sherbrooke connectivity Imaging Lab (SCIL), University Hospital of Sherbrooke, Canada.

Contact: Laurent Petit -Groupe d'Imagerie Neurofonctionnelle (GIN), UMR5296-CNRS-CEA, laurent.petit@u-bordeaux.fr

Maxime Descoteaux is Associate Professor at the Département d’Informatique, Faculté des Sciences, CHU de Sherbrooke (CHUS), Québec, Canada.
He is also the scientific director of the image analysis and visualization plateform (PAVI) of the Centre de Recherche CHUS.
He did a post-doctorat at NeuroSpin under the supervision of Cyril Poupon. He also obtained a PhD in Computer Science at INRIA Sophia Antipolis - Mediterranée, supervised by R. Deriche.
In this talk, he will introduce the concept of bundle analytics. As of today, there exist very little computational tools to process tractograms, fiber bundles or streamlines, in their specific space, i.e. the space of streamlines. Just as surface meshes can improve on the precision of registration and segmentation in the imageworld for T1 images for example, streamlines and their 3D point-based representation can lead to new powerful techniques for tractography analysis. He will show examples of fiber bundle clustering, automatic bundle recognition/segmentation, outlier rejection and atlas construction based on bundle analytics.