基于动态全变差的动态磁共振图像重建

摘要

动态核磁共振(dMRI)成像技术要在极短的扫描时间内获取高时空精度的MR图像,目前仍然是一个难点.文中提出了一种新的基于动态全变差技术(DTV)的压缩感知(CS)动态磁共振图像重建方案,该方法充分利用dMRI数据在时空域的稀疏性.文中算法采用加权最小二乘算法解决重建速度过慢的问题.论文将该算法与目前比较先进的两种算法TV及k-t SLR做比较,实验数据来源于临床脑部电影数据集.实验结果显示该方法在重建精度上处于领先水平.%Dynamic magnetic resonance imaging (DMRI) needs to obtain MR images with high spatiotemporal resolution in a short scan time,which is still a challenging problem.This paper proposes an innovative technologies called dynamic total variation for comprehensive sensing (CS) DMRI reconstruction.The proposed method is used to exploit the sparsity on both spatial and temporal domains.In addition,it designs an accelerated weighted least squares algorithm to solve the challenging reconstruction problem.The proposed method is compared with TV and k-t SLR methods on the clincal brain cine dataset.The experimental results show that the method significantly outperforms these methods in terms of reconstruction accuracy.