MRI图像重建网格化算法的研究进展

磁共振成像常常是在空间频率域不同间隔地采集数据的。为了利用快速傅里叶变换进行图像重建 ,采集到的 MRI数据必须用网格化算法转换到等间隔的直线网格上。目前有 6种不同的网格化算法 ,其中 Jackson的 Gridding法和Rosenfeld的矩阵反演法已成为该领域的主流方法 ,但两者都不太完善。具有创新思路的网格化算法仍有待于学术界进一步的开拓

Overview of gridding method in MRI reconstruction

Magnetic resonance imaging data are often coll ec ted along non-rectilinear k-space trajectories.In order to take the advantage of fast Foruier transform in image reconstruction,the collected data must be res ampled onto a rectilinear grid using gridding method.Up to six different griddin g strategies are described and criticized.The ones proposed by Jackson et al and by Rosenfeld are widely accepted as first choices in regridding.The both have n ot,however,been as perfect as expected yet.New strategies for gridding are high ly disired.