A Slicing-Based Coherence Measure for Clusters of DTI Integral Curves

We present a slicing-based coherence measure for clusters of DTI integral curves. For a given cluster, we probe samples from the cluster by slicing it with a plane at regularly spaced locations parametrized by curve arc lengths. Then we compute a stability measure based on the spatial relations between the projections of the curve points in individual slices and their change across the slices. We demonstrate its use in refining agglomerative hierarchical clustering results of DTI curves that correspond to neural pathways. Expert evaluation shows that refinement based on our measure can lead to improvement of clustering that is not possible directly by using standard methods.