Tractography segmentation using a hierarchical Dirichlet processes mixture model

►A new nonparametric Bayesian framework to cluster white matter fiber tracts into bundles using a hierarchical Dirichlet processes mixture model; ►The number of clusters is automatically learned driven by data with a Dirichlet process prior instead of being manually specified; ►The models of bundles learned from the training set are used as priors to cluster fibers of new subjects for comparison across subjects; ►The proposed approach has a lower space complexity compared with distance-based clustering methods since it does not require computing pairwise distances.