结合导航回波与压缩感知进行运动伪影矫正

Motion Artifacts Correction in MRI with Navigator Echo Combined with Compressed Sensing

在一定幅度范围内,导航回波可有效追踪刚体运动,用于矫正k空间数据,从而降低运动伪影对图像质量的影响.然而导航回波技术无法很好地矫正大幅度运动以及非刚体运动导致的图像伪影,运动时刻采集到的k空间数据只能舍弃.压缩感知通过非线性规划法,对欠采样数据进行重建,能恢复出原始信号.该文采用伪随机的方式进行数据采集,结合导航回波技术,用压缩感知对未受运动影响的数据进行图像重建,从而减少运动伪影对图像的干扰.该研究为运动伪影的矫正提供了一种新思路.

In MRI, rigid-body motion with limited amplitude can be tracked effectively by using navigator echo, and then corrected with the information provided by the navigator echo. However, tracking and correction of large and/or non-rigid motion artifact by means of navigator echo could be problematic. Images with a sparse representation in a given domain can be accurately reconstructed from undersampled datasets by using nonlinear reconstruction algorithms, such as compressed sensing. In this study, we developed an imaging technique using pseudo-randomk-space trajectory for acquisition combined with motion tracking by navigator echo. After discarding the motion corrupted lines ink-space, the randomly undersampled dataset was reconstructed using compressed sensing. Experimental results demonstrated that good image quality can be obtained with this technique even in the presence of large non-rigid motions.