Evaluation of Local Filter Approaches for Diffusion Tensor-Based Fiber Tracking

Diffusion tensor imaging (DTI) allows investigating white matter structures in vivo which is of great interest for various applications in neuroanatomy and neurosurgery. In order to reconstruct white matter tracts from DTI data, fiber tracking approaches are used which are commonly based on streamline propagation techniques. In order to provide more reliable tracking results, local regularization filters were presented previously, which can be applied during the fiber tracking process and move forward as the fibers are constructed. In this work, Gaussian and least-squares local filters are evaluated and compared against higher order integration schemes. The results provide a more differentiated view on the prospects and limitations of local regularization filters.