基于薄板样条的MRI图像与脑图谱的配准方法

为了将CT、MRI、PET或SPECT等断层扫描图像用于疾病的辅助诊断、放射治疗、手术计划和引导 ,就必须知道图像中感兴趣区 (ROI)是什么解剖组织 ,即解决医学图像的解剖标识问题。医生通常是从解剖书籍、图谱及自身经验来对ROI做出判断。这些书籍和图谱往往给出的是文字描述和有限数目的 ,固定位置和方向的断层图片 ,很难与患者的实际图像联系起来。对于缺乏临床经验的医生来说尤为困难。数字化 3D人脑解剖图谱使医生对人脑深部组织全方位可视化。因此可以将其通过一定的空间变换 ,与MR体积数据集中的ROI进行比较 ,从而得到ROI的解剖标识。但是 ,没有两个人的大脑是完全一样的。人脑的解剖个体差异较大 ,这就要求利用非线性变形的方法做解剖标识。本研究介绍通过薄板样条变换用Talirach脑图谱对MR图像做解剖标识的方法。

REGISTRATION BETWEEN MR IMAGE AND BRAIN ATLAS BASED ON THIN-PLATE SPLINES

For clinical applications of medical tomography images, e.g.CT,MRI,it is necessary to label region of interest anatomically. Since there is no topological equivalence between two different brains, nonlinear transformation is needed for the registration between MR images and Talairach atlas. In this paper, we described a thin-plate spline based method used for labeling of anatomic structures in MR images. Even with manual selection of landmarks, the results for datasets of MNI, VHP and two Chinese subjects are satisfied.The method could be used for inter-subject registration.