I already searched the topics here but couldn't figure out clearly what to do so I'm asking here. I have thousands of fiber tracts (the outputs of a tractography algorithm on brain diffusion MRI) that I'd like to cluster them. Each tract is essentially a series of points in 3D Cartesian space. Not all the tracts have the same length and start and end point. How can I extend the DTW to compute the similarity between fiber tracts?