k-Space Deep Learning for Accelerated MRI
k-Space Deep Learning for Accelerated MRI
复制标题
DOI：
10.1109/tmi.2019.2927101
复制
发表时间：
2020-02-01
影响因子：
10.6
通讯作者：
Ye, Jong Chul
中科院分区：
文献类型：
--
作者：
Han, Yoseob;Sunwoo, Leonard;Ye, Jong Chul
基于零化滤波的低阶Hankel矩阵方法(ALOHA)是目前最先进的压缩感知方法之一，它利用低阶Hankel矩阵补全来直接内插丢失的k-空间数据。ALOHA算法的成功要归功于k-空间域的简洁信号表示，这要归功于k-空间域的结构化低秩性和图像域稀疏性之间的对偶性。受最近发现的卷积神经网络与Hankel矩阵分解相联系的数学发现的启发，本文提出了一种完全数据驱动的k-空间内插深度学习算法。我们的网络也可以很容易地应用于非笛卡尔k空间轨迹，只需增加一个额外的重新叠加层。大量的数值实验表明，该深度学习方法的性能始终优于现有的图像域深度学习方法。
The annihilating filter-based low-rank Hankel matrix approach (ALOHA) is one of the state-of-the-art compressed sensing approaches that directly interpolates the missing k-space data using low-rank Hankelmatrix completion. The success of ALOHA is due to the concise signal representation in the k-space domain, thanks to the duality between structured low-rankness in the k-space domain and the image domain sparsity. Inspired by the recent mathematical discovery that links convolutional neural networks to Hankel matrix decomposition using data-driven framelet basis, here we propose a fully data-driven deep learning algorithm for k-space interpolation. Our network can be also easily applied to non-Cartesian k-space trajectories by simply adding an additional regridding layer. Extensive numerical experiments show that the proposed deep learningmethod consistently outperforms the existing image-domain deep learning approaches.