A model-guided string-based approach to white matter fiber-bundles extraction

Abstract

In this paper we present a new model-guided approach to extracting anatomically plausible White Matter fiber-bundles from the high number of streamlines generated by tractography algorithms. Our approach is based on: (i) an approximate shape model of certain fiberbundles constructed by an expert operator; (ii) a particular string representation of fibers; (iii) a new string similarity metric. It transforms the fiber-bundles of both the model and the tractography streamlines into strings and uses the string similarity metric for comparison and extraction tasks.