Optimized White Matter Fiber Reconstruction with B-Spline Curve and Evolutionary Computation

Diffusion tensor tractography can reconstruct white matter fiber bundle non-invasively. In this paper, a novel fiber tracking algorithm is proposed based on an adaption of evolutionary computation. Given a pair of ROIs, the proposed algorithm can generate globally optimized fiber pathways in terms of fiber geometric smoothness and consistency with the tensor field with the B-spline curve model. The experiments demonstrate that the fiber tracking algorithm we proposed can reconstruct white matter fiber trajectories faithfully for both synthetic and in vivo DTI data. Compared to existing methods, the new tracking algorithm incorporates a global constraint, thus making it insusceptible to image noise and other local artifacts.