白质纤维重建过程中像素内交叉问题解决方法  

Method for Intra-Voxel Fiber Crossing Problem During White Matter Fiber Reconstruction

作　　者：王明时[1] 高伟[1] 赵欣[1] 路会生[1] 刘海婴[2] 王力群 李德彪[4] 

机构地区：[1]天津大学精密仪器与光电子工程学院,天津300072 [2]明尼苏达大学放射学院,美国明尼苏达州55455 [3]东京电机大学超导应用研究所高技术研究中心 [4]西北大学生物医学工程系,美国芝加哥60611

出　　处：《天津大学学报》2006年第3期354-359,共6页Journal of Tianjin University(Science and Technology)

基　　金：天津市科技发展计划资助项目(04310211).

摘　　要：由部分体积效应引起的像素内交叉现象一直是利用扩散张量图像(DT-MRI)数据重建脑内白质纤维过程中的一个主要问题,而区分交叉区域和非交叉区域是解决这一问题的难点和关键．为此,提出了一种新的方法试图定义“交叉像素”和“非交叉像素”,并设计了一种改进的优化算法来恢复交叉像素内的多纤维取向信息,从而使纤维跟踪过程在交叉区域和非交叉区域分别采取各自最优的方向依据．模拟数据和真实扩散张量图像数据跟踪实验显示,该方法首先判断了交叉区域,然后纤维沿恢复后的方向通过了交叉区域,从而解决了白质纤维重建过程中的像素内纤维交叉问题,提高了跟踪精度,并减少了计算量．Intra-voxel fiber crossing induced by partial volume effect has long been a key problem of white matter fiber reconstruction in brain using diffusion tensor magnetic resonance imaging （DT-MRI） data. While the classification of crossing areas and non-crossing areas is crucial and remains a challenge. In this paper, a new method to define ＂crossing voxels＂ and ＂non-crossing voxels＂ is proposed, and a modified regularization method to restore the underlying multiple fiber orientations in crossing voxels is presented. Hence fiber trac-king in crossing and non-crossing areas can adopt optimum direction measurements respectively. Both synthetic and real DT-MRI data experiments show that this method is effective to judge crossing areas and then fibers follow the re- stored direction in the classified crossing areas to resolve the intra-voxel fiber crossing problem during white matter fiber recomstruotion, which improves tracking accuray and simultaneously reduces computation accurary.

关 键 词：像素内交叉 扩散张量 白质纤维重建 部分体积效应 

分 类 号：R445.2[医药卫生—影像医学与核医学]