Rapid compressed sensing reconstruction of 3D non-Cartesian MRI.

Rapid compressed sensing reconstruction of 3D non-Cartesian MRI.
复制标题

DOI:
10.1002/mrm.26928
复制
发表时间:
2018-05
影响因子:
3.3
通讯作者:
Nishimura DG
Nishimura DG
中科院分区:
医学3区
文献类型:
--
作者:
Baron CA;Dwork N;Pauly JM;Nishimura DG

文献摘要

参考文献

被引文献

相似文献

传统的非笛卡尔压缩感知每次迭代都需要多次非均匀傅里叶变换(NUFFT),这在计算上是昂贵的。因此,耗时的重建减缓了欠采样 3D 非笛卡尔采集在临床方案中的采用。在这项工作中,我们研究了几种在不牺牲准确性的情况下最大限度地减少重建时间的方法。重建问题可以重新格式化,以利用每次迭代都会评估的矩阵的 Toeplitz 结构,但它需要比 NUFFT 严格要求的更大的过采样。因此,我们研究了两种方法针对不同 NUFFT 内核大小以及 GPU 和 CPU 实现的过采样的相对速度。其次,我们引入一种通过估计图像支持来最小化矩阵大小的方法。最后,密度补偿权重被用作预处理矩阵来提高收敛性,但这会增加噪声。我们提出了一种更通用的预处理方法,可以在精度和收敛速度之间进行权衡。使用 GPU 时,Toeplitz 方法对于所有实际参数都更快。其次,我们发现,正确考虑图像支持可以防止锯齿错误,同时对重建时间的影响最小。第三,所提出的预处理方案将收敛速度提高了一个数量级,而对噪声的影响可以忽略不计。通过所提出的方法,使用实用的计算机资源,具有临床相关重建时间(< 2 分钟)的 3D 非笛卡尔压缩感知是可行的。
Conventional non-Cartesian compressed sensing requires multiple nonuniform Fourier transforms (NUFFTs) every iteration, which is computationally expensive. Accordingly, time-consuming reconstructions have slowed the adoption of undersampled 3D non-Cartesian acquisitions into clinical protocols. In this work we investigate several approaches to minimize reconstruction times without sacrificing accuracy. The reconstruction problem can be reformatted to exploit the Toeplitz structure of matrices that are evaluated every iteration, but it requires larger oversampling than what is strictly required by NUFFT. Accordingly, we investigate relative speeds of the two approaches for various NUFFT kernel sizes and oversampling for both GPU and CPU implementations. Second, we introduce a method to minimize matrix sizes by estimating the image support. Finally, density compensation weights have been used as a preconditioning matrix to improve convergence, but this increases noise. We propose a more general approach to preconditioning that allows a trade-off between accuracy and convergence speed. When using a GPU, the Toeplitz approach was faster for all practical parameters. Second, it was found that properly accounting for image support can prevent aliasing errors with minimal impact on reconstruction time. Third, the proposed preconditioning scheme improved convergence rates by an order of magnitude with negligible impact on noise. With the proposed methods, 3D non-Cartesian compressed sensing with clinically relevant reconstruction times (< 2 min) is feasible using practical computer resources.
DOI: 10.1002/mrm.24751
发表时间: 2014-03
影响因子: 3.3
作者:
Uecker, Martin;Lai, Peng;Murphy, Mark J.;Virtue, Patrick;Elad, Michael;Pauly, John M.;Vasanawala, Shreyas S.;Lustig, Michael
通讯作者: Lustig, Michael
DOI: 10.1137/080716542
发表时间: 2009-01-01
影响因子: 2.1
作者:
Beck, Amir;Teboulle, Marc
通讯作者: Teboulle, Marc
DOI: 10.1109/tsp.2005.853152
发表时间: 2005-09-01
影响因子: 5.4
作者:
Fessler, JA;Lee, S;Noll, DC
通讯作者: Noll, DC
DOI: 10.1118/1.4936098
发表时间: 2015-12-01
期刊: MEDICAL PHYSICS
影响因子: 3.8
作者:
Tao, Shengzhen;Trzasko, Joshua D.;Bernstein, Matt A.
通讯作者: Bernstein, Matt A.
DOI: 10.1016/j.jmr.2014.01.016
发表时间: 2014-05
影响因子: 2.2
作者:
Yang, Zhili;Jacob, Mathews
通讯作者: Jacob, Mathews