磁共振弥散加权成像涡流失真系统性矫正算法研究

Systematic Correction Algorithm for Eddy Current Distortion in Diffusion Weighted Imaging of Magnetic Resonance

针对基于EPI扫描序列的弥散加权成像技术中存在的涡流失真现象,给出一种基于图像配准的系统性失真矫正方法.矫正过程以非弥散加权图像为参考图像,利用二值蒙版分割图像消除高b值图像脑脊液区域对比度差异影响,在迭代互相关ICC算法基础上结合相邻层的变形参数进行平滑得到系统性配准参数.比较实验表明,结合相邻层信息能减少两两配准带来的矫正误差,得到更好的矫正效果.

The paper provided a kind of systematic distortion correction algorithm based on image registration for the eddy current distortion in diffusion weighted imaging technology on the basis of EPI sequence. The correction process took non-diffusion weighted images as reference images, eliminated the influence of contrast differences in the cerebral spinal fluid area with high b value by using binary mask to segment images, and obtained systematic registration parameters by smoothing the deformation parameters of adjacent layers based on the iterative cross correlation ICC algorithm. The experiment shows that the information combination of adjacent layers can reduce the correction errors which are brought by pairwise registration, and obtain better correction effects.