基于向量选择的神经纤维跟踪改进算法

为了使跟踪出的神经纤维更加平滑,本研究在基于向量选择跟踪算法的基础上,提出了一种改进方法。首先将向量选择的标准细化,区分出与前步纤维路经方向大致相反的体素;然后将此类体素的距离权重系数进行调整,将反方向贡献转化为正向贡献实现平滑跟踪。对比实验结果表明:本研究的改进算法能够有效去除纤维跟踪过程中出现的抖动,使纤维路径延长,更好的表现神经纤维的解剖特点和脑白质区域的连通性。

Advanced Fiber Tracking Algorithm by Vector Selection Criterion in DTI images

To further smooth the neural fibers,a fiber tracking algorithm based on vector selection was proposed.First,the vector selection criterion was specialized to identify those voxels whose directions are approximately opposite to the previous fiber tracking.Then the distance-weighted coefficients of such voxels was adjusted to transform the reverse contribution into the obverse for smoothing fibers.The contrastive experiment result validates that the proposed algorithm is efficiently able to eliminate the zigzags in the fiber tracking and elongate the fiber pathway for better reflecting the anatomic characteristics of neural fiber as well as the connection in white matter.