Systematic evaluation of iterative deep neural networks for fast parallel MRI reconstruction with sensitivity‐weighted coil combination
Open Access
- 10 June 2021
- journal article
- research article
- Published by Wiley in Magnetic Resonance in Medicine
- Vol. 86 (4), 1859-1872
- https://doi.org/10.1002/mrm.28827
Abstract
Purpose To systematically investigate the influence of various data consistency layers and regularization networks with respect to variations in the training and test data domain, for sensitivity-encoded accelerated parallel MR image reconstruction. Theory and Methods Magnetic resonance (MR) image reconstruction is formulated as a learned unrolled optimization scheme with a down-up network as regularization and varying data consistency layers. The proposed networks are compared to other state-of-the-art approaches on the publicly available fastMRI knee and neuro dataset and tested for stability across different training configurations regarding anatomy and number of training samples. Results Data consistency layers and expressive regularization networks, such as the proposed down-up networks, form the cornerstone for robust MR image reconstruction. Physics-based reconstruction networks outperform post-processing methods substantially for R = 4 in all cases and for R = 8 when the training and test data are aligned. At R = 8, aligning training and test data is more important than architectural choices. Conclusion In this work, we study how dataset sizes affect single-anatomy and cross-anatomy training of neural networks for MRI reconstruction. The study provides insights into the robustness, properties, and acceleration limits of state-of-the-art networks, and our proposed down-up networks. These key insights provide essential aspects to successfully translate learning-based MRI reconstruction to clinical practice, where we are confronted with limited datasets and various imaged anatomies.Keywords
This publication has 50 references indexed in Scilit:
- Low-Rank Modeling of Local $k$-Space Neighborhoods (LORAKS) for Constrained MRIIEEE Transactions on Medical Imaging, 2013
- Calibrationless parallel imaging reconstruction based on structured low-rank matrix completionMagnetic Resonance in Medicine, 2013
- ESPIRiT—an eigenvalue approach to autocalibrating parallel MRI: Where SENSE meets GRAPPAMagnetic Resonance in Medicine, 2013
- Parallel imaging with nonlinear reconstruction using variational penaltiesMagnetic Resonance in Medicine, 2011
- Fields of ExpertsInternational Journal of Computer Vision, 2009
- Sparse MRI: The application of compressed sensing for rapid MR imagingMagnetic Resonance in Medicine, 2007
- Undersampled radial MRI with multiple coils. Iterative image reconstruction using a total variation constraintMagnetic Resonance in Medicine, 2007
- Matrix description of general motion correction applied to multishot imagesMagnetic Resonance in Medicine, 2005
- Generalized autocalibrating partially parallel acquisitions (GRAPPA)Magnetic Resonance in Medicine, 2002
- SENSE: Sensitivity encoding for fast MRIMagnetic Resonance in Medicine, 1999