Self-Supervised Learning of Physics-Guided Reconstruction Neural Networks without Fully-Sampled Reference Data

Abstract

Purpose: To develop a strategy for training a physics-guided MRI reconstruction neural network without a database of fully-sampled datasets. Theory and Methods: Self-supervised learning via data under-sampling (SSDU) for physics-guided deep learning (DL) reconstruction partitions available measurements into two disjoint sets, one of which is used in the data consistency units in the unrolled network and the other is used to define the loss for training. The proposed training without fully-sampled data is compared to fully-supervised training with ground-truth data, as well as conventional compressed sensing and parallel imaging methods using the publicly available fastMRI knee database. The same physics-guided neural network is used for both proposed SSDU and supervised training. The SSDU training is also applied to prospectively 2-fold accelerated high-resolution brain datasets at different acceleration rates, and compared to parallel imaging. Results: Results on five different knee sequences at acceleration rate of 4 shows that proposed self-supervised approach performs closely with supervised learning, while significantly outperforming conventional compressed sensing and parallel imaging, as characterized by quantitative metrics and a clinical reader study. The results on prospectively sub-sampled brain datasets, where supervised learning cannot be employed due to lack of ground-truth reference, show that the proposed self-supervised approach successfully perform reconstruction at high acceleration rates (4, 6 and 8). Image readings indicate improved visual reconstruction quality with the proposed approach compared to parallel imaging at acquisition acceleration. Conclusion: The proposed SSDU approach allows training of physics-guided DL-MRI reconstruction without fully-sampled data, while achieving comparable results with supervised DL-MRI trained on fully-sampled data.

This work is an extension of our previous work arXiv:1910.09116

References in corpus (7)

• RARE: Image Reconstruction using Deep Priors Learned without Ground Truth
• Dense Recurrent Neural Networks for Accelerated MRI: History-Cognizant Unrolling of Optimization Algorithms
• Wasserstein GANs for MR Imaging: from Paired to Unpaired Training
• Self-Supervised Physics-Based Deep Learning MRI Reconstruction Without Fully-Sampled Data
• Brain MRI super-resolution using 3D generative adversarial networks
• Deep MRI Reconstruction: Unrolled Optimization Algorithms Meet Neural Networks
• Unsupervised Deep Basis Pursuit: Learning inverse problems without ground-truth data

Cited by in corpus (45)

• Results of the 2020 fastMRI Challenge for Machine Learning MR Image Reconstruction
• Adaptive Diffusion Priors for Accelerated MRI Reconstruction
• Noise2Inverse: Self-supervised deep convolutional denoising for tomography
• Dense Recurrent Neural Networks for Accelerated MRI: History-Cognizant Unrolling of Optimization Algorithms
• Unsupervised Deep Learning Methods for Biological Image Reconstruction and Enhancement
• Multi-Mask Self-Supervised Learning for Physics-Guided Neural Networks in Highly Accelerated MRI
• Self-Supervised Coordinate Projection Network for Sparse-View Computed Tomography
• SGD-Net: Efficient Model-Based Deep Learning with Theoretical Guarantees
• Zero-Shot Self-Supervised Learning for MRI Reconstruction
• Self-Supervised Physics-Guided Deep Learning Reconstruction For High-Resolution 3D LGE CMR
• ENSURE: A General Approach for Unsupervised Training of Deep Image Reconstruction Algorithms
• Zero-DeepSub: Zero-Shot Deep Subspace Reconstruction for Rapid Multiparametric Quantitative MRI Using 3D-QALAS
• CoIL: Coordinate-based Internal Learning for Imaging Inverse Problems
• Data and Physics driven Deep Learning Models for Fast MRI Reconstruction: Fundamentals and Methodologies
• Self-Supervised Learning for Improved Calibrationless Radial MRI with NLINV-Net
• Scan-specific Self-supervised Bayesian Deep Non-linear Inversion for Undersampled MRI Reconstruction
• Physics-informed self-supervised deep learning reconstruction for accelerated first-pass perfusion cardiac MRI
• A Tutorial on MRI Reconstruction: From Modern Methods to Clinical Implications
• Learning Task-Specific Strategies for Accelerated MRI
• Robust multi-coil MRI reconstruction via self-supervised denoising
• Validation and Generalizability of Self-Supervised Image Reconstruction Methods for Undersampled MRI
• Semi-Supervised Learning of Mutually Accelerated MRI Synthesis without Fully-Sampled Ground Truths
• SHFormer: Dynamic Spectral Filtering Convolutional Neural Network and High-pass Kernel Generation Transformer for Adaptive MRI Reconstruction
• On Instabilities of Conventional Multi-Coil MRI Reconstruction to Small Adverserial Perturbations
• mcLARO: Multi-Contrast Learned Acquisition and Reconstruction Optimization for simultaneous quantitative multi-parametric mapping
• Self-supervised feature learning for cardiac Cine MR image reconstruction
• Zero-Shot Adaptation for Approximate Posterior Sampling of Diffusion Models in Inverse Problems
• Self-Supervised Adversarial Diffusion Models for Fast MRI Reconstruction
• End-to-End AI-based MRI Reconstruction and Lesion Detection Pipeline for Evaluation of Deep Learning Image Reconstruction
• NoSENSE: Learned unrolled cardiac MRI reconstruction without explicit sensitivity maps
• Fast MRI Reconstruction Using Deep Learning-based Compressed Sensing: A Systematic Review
• Advancing MRI Reconstruction: A Systematic Review of Deep Learning and Compressed Sensing Integration
• Unsupervised Learning From Incomplete Measurements for Inverse Problems
• Improved Supervised Training of Physics-Guided Deep Learning Image Reconstruction with Multi-Masking
• Adaptive Local Neighborhood-based Neural Networks for MR Image Reconstruction from Undersampled Data
• Learned Primal Dual Splitting for Self-Supervised Noise-Adaptive MRI Reconstruction
• MRI Recovery with Self-Calibrated Denoisers without Fully-Sampled Data
• Generative Priors for MRI Reconstruction Trained from Magnitude-Only Images Using Phase Augmentation
• Noise2Inpaint: Learning Referenceless Denoising by Inpainting Unrolling
• Self-Supervised Joint Reconstruction and Denoising of T2-Weighted PROPELLER MRI of the Lungs at 0.55T
• Magnetic Resonance Parameter Mapping using Self-supervised Deep Learning with Model Reinforcement
• Deep Multi-view Depth Estimation with Predicted Uncertainty
• Self-Supervised Learning for MRI Reconstruction with a Parallel Network Training Framework
• Single Image Compressed Sensing MRI via a Self-Supervised Deep Denoising Approach
• Self-Supervised Noise Adaptive MRI Denoising via Repetition to Repetition (Rep2Rep) Learning