Jiangsu Key Laboratory of Meteorological Observation and Information Processing, Nanjing 210044, China; Jiangsu Technology and Engineering Center of Meteorological Sensor Network, Nanjing 210044, China; School of Electronic and Information Engineering, Nanjing 210044, China; Nanjing University of Information Science and Technology, Nanjing 210044, China.
Jiangsu Key Laboratory of Meteorological Observation and Information Processing, Nanjing 210044, China; Jiangsu Technology and Engineering Center of Meteorological Sensor Network, Nanjing 210044, China; School of Electronic and Information Engineering, Nanjing 210044, China; Nanjing University of Information Science and Technology, Nanjing 210044, China.
Magn Reson Imaging. 2019 Nov;63:93-104. doi: 10.1016/j.mri.2019.07.014. Epub 2019 Jul 27.
Magnetic resonance imaging (MRI) reconstruction is an active inverse problem which can be addressed by conventional compressed sensing (CS) MRI algorithms that exploit the sparse nature of MRI in an iterative optimization-based manner. However, two main drawbacks of iterative optimization-based CSMRI methods are time-consuming and are limited in model capacity. Meanwhile, one main challenge for recent deep learning-based CSMRI is the trade-off between model performance and network size. To address the above issues, we develop a new multi-scale dilated network for MRI reconstruction with high speed and outstanding performance. Comparing to convolutional kernels with same receptive fields, dilated convolutions reduce network parameters with smaller kernels and expand receptive fields of kernels to obtain almost same information. To maintain the abundance of features, we present global and local residual learnings to extract more image edges and details. Then we utilize concatenation layers to fuse multi-scale features and residual learnings for better reconstruction. Compared with several non-deep and deep learning CSMRI algorithms, the proposed method yields better reconstruction accuracy and noticeable visual improvements. In addition, we perform the noisy setting to verify the model stability, and then extend the proposed model on a MRI super-resolution task.
磁共振成像（MRI）重建是一个活跃的反问题，可以通过传统的压缩感知（CS）MRI 算法来解决，这些算法以迭代优化的方式利用 MRI 的稀疏性。然而，基于迭代优化的 CS MRI 方法主要有两个缺点：耗时和模型容量有限。同时，基于最近的深度学习的 CS MRI 的一个主要挑战是模型性能和网络大小之间的权衡。为了解决上述问题，我们开发了一种新的用于 MRI 重建的多尺度扩张网络，具有高速和出色的性能。与具有相同感受野的卷积核相比，扩张卷积通过使用较小的核减少网络参数，并扩展核的感受野以获得几乎相同的信息。为了保持特征的丰富性，我们提出了全局和局部残差学习，以提取更多的图像边缘和细节。然后，我们利用串联层融合多尺度特征和残差学习，以实现更好的重建。与几种非深度学习和深度学习的 CS MRI 算法相比，所提出的方法具有更好的重建精度和显著的视觉改进。此外，我们还进行了噪声设置以验证模型稳定性，然后将所提出的模型扩展到 MRI 超分辨率任务。
