Abstract:

Probabilistic fiber tracking algorithm only uses the maximum probability to track fibers,and ignores some orientations which have big probabilities. And the speed of calculation is slow. So a fast probabilistic fiber tracking algorithm based on the threshold is proposed. This paper sets the threshold which can find more crossing and branching fibers. Simplifying the parameters of calculation can improve the speed without affecting the effect of fiber tracking. Experimental result shows that this algorithm can reflect the distribution of the neural fibers in the cerebral white matter and reduces the time of calculation compared with probabilistic fiber tracking algorithm.

摘要：

概率跟踪算法仅对行走概率最大的方向进行跟踪,忽略了纤维走向概率较大的方向,且运算速度较慢。为此,提出一种基于阈值的快速概率跟踪算法。设定纤维走向的概率阈值,以找到更多交叉和分叉的纤维,在不影响纤维跟踪效果的情况下,对计算参数进行简化,从而提高运算速度。实验结果表明,与概率跟踪算法相比,该算法能更好地反映脑白质内神经纤维的分布情况,且缩短运算时间。

关键词: 纤维跟踪, 概率跟踪算法, 扩散张量, 各向异性, 阈值

CLC Number: 

• TP391