HAMLET: Hierarchical Harmonic Filters for Learning Tracts from Diffusion\n MRI
No TL;DR found
Abstract
In this work we propose HAMLET, a novel tract learning algorithm, which,\nafter training, maps raw diffusion weighted MRI directly onto an image which\nsimultaneously indicates tract direction and tract presence. The automatic\nlearning of fiber tracts based on diffusion MRI data is a rather new idea,\nwhich tries to overcome limitations of atlas-based techniques. HAMLET takes a\nsuch an approach. Unlike the current trend in machine learning, HAMLET has only\na small number of free parameters HAMLET is based on spherical tensor algebra\nwhich allows a translation and rotation covariant treatment of the problem.\nHAMLET is based on a repeated application of convolutions and non-linearities,\nwhich all respect the rotation covariance. The intrinsic treatment of such\nbasic image transformations in HAMLET allows the training and generalization of\nthe algorithm without any additional data augmentation. We demonstrate the\nperformance of our approach for twelve prominent bundles, and show that the\nobtained tract estimates are robust and reliable. It is also shown that the\nlearned models are portable from one sequence to another.\n