【作者】 刘卫华；

【导师】 姜杉；

【作者基本信息】 天津大学 ， 机械工程， 2012， 硕士

【摘要】 扩散张量成像(DTI)作为磁共振成像(MRI)研究上的一项重大技术进展和突破,是目前唯一可以无创地、清晰地显示人脑内部组织的纤维结构的技术。它通过对大脑白质区域水分子各向异性扩散信息的表达来研究大脑白质纤维的走向和三维结构信息。本课题基于人脑磁共振扩散张量图像数据集,对扩散张量数据集的处理、显示方法和白质纤维追踪算法进行了研究,主要研究内容、成果如下:首先,在对扩散张量成像原理进行充分理解的基础上,对扩散张量的数据集的可视化方法进行了研究,包括颜色编码法、图元显示法、纤维追踪法和体显示法。其中重点对图元显示法中不同的图元对不同区域、不同大小的DTI数据集进行了显示,总结了不同图元各自的优缺点,并采用实际数据进行方法性验证。其次,对白质束纤维追踪的算法进行了深入研究,在总结前人所提出的追踪算法的优缺点基础上,分析这些算法不能解决纤维的交叉分支问题的原因,而快速行进算法能够解决该问题,但伴随而来的是假性阳支的出现,所以本论文在前人的基础之上,提出了改进的快速行进算法,通过不同体素之间的演化,定义不同的速度函数,使得新算法具有可靠的理论基础。最后,提出将三线性插值算法运用到白质束追踪当中。通过VTK(Visual Toolkit)软件包进行仿真,实现了对该算法在纤维追踪上的运用,并且得到了较好的追踪结果。为后续进行DTI数据集仿真奠定了一个良好的理论基础。更多还原

【Abstract】 Diffusion Tensor Imaging (DTI) which as one of the major breakthrough of the research on Magnetic Resonance Imaging (MRI) is the unique noninvasive in vivo imaging technology to generate fiber-tract trajectories within the white matter of human brain. Through the showing of water molecules in brain white matter area, DTI is used to research the fiber direction of the brain nerve and its three dimensions configuration.This thesis bases on the MRI data set has accomplish the research on the processing and rending method of DTI data set and white matter fiber tracking algorithm, the main work and achievement can be concluded as following:Firstly, on the basis of fully understanding of the DTI principle, this thesis has studied the method of visualization of DTI data set, including Color coding, Glyphing, White matter tractography and Volume Rending. The key work of this part is the rending of DTI data set of different region and size with different glyphs, and then the advantages and disadvantages of these glyphs are concluded. Also, the conclusions have been confirmed by experimental results of the real DTI data set.Secondly, this thesis has researched deeply the white matter fiber tracking algorithm. Based on reviewing the advantages of the algorithms introduced by predecessors, the issue that those algorithms can’t solve the problem of the crossing and branching of white matter fibers is analysed, yet Fast Marching Method (FMM) can solve the problem. But FMM gets a lot of false pathway, so this thesis has proposed an improved FMM for FT, which defines different speed function according to the different evolving between the voxels, so the improved FMM has the reliable theory foundation.Finally, we proposed the tri-linear interpolation algorithm for fiber tracking (FT). And we have achieved the white matter fibers simulation with VTK (Visual Toolkit), the result is coincide with the real fibers by and large. The work can be the theoretical basis of future research.更多还原