Dense Recurrent Neural Networks for Accelerated MRI: History-Cognizant Unrolling of Optimization Algorithms

Abstract

Inverse problems for accelerated MRI typically incorporate domain-specific knowledge about the forward encoding operator in a regularized reconstruction framework. Recently physics-driven deep learning (DL) methods have been proposed to use neural networks for data-driven regularization. These methods unroll iterative optimization algorithms to solve the inverse problem objective function, by alternating between domain-specific data consistency and data-driven regularization via neural networks. The whole unrolled network is then trained end-to-end to learn the parameters of the network. Due to simplicity of data consistency updates with gradient descent steps, proximal gradient descent (PGD) is a common approach to unroll physics-driven DL reconstruction methods. However, PGD methods have slow convergence rates, necessitating a higher number of unrolled iterations, leading to memory issues in training and slower reconstruction times in testing. Inspired by efficient variants of PGD methods that use a history of the previous iterates, we propose a history-cognizant unrolling of the optimization algorithm with dense connections across iterations for improved performance. In our approach, the gradient descent steps are calculated at a trainable combination of the outputs of all the previous regularization units. We also apply this idea to unrolling variable splitting methods with quadratic relaxation. Our results in reconstruction of the fastMRI knee dataset show that the proposed history-cognizant approach reduces residual aliasing artifacts compared to its conventional unrolled counterpart without requiring extra computational power or increasing reconstruction time.

References in corpus (6)

• Self-Supervised Learning of Physics-Guided Reconstruction Neural Networks without Fully-Sampled Reference Data
• SIMBA: Scalable Inversion in Optical Tomography using Deep Denoising Priors
• Recurrent Generative Adversarial Networks for Proximal Learning and Automated Compressive Image Recovery
• Deep Learning Methods for Parallel Magnetic Resonance Image Reconstruction
• Unsupervised Deep Basis Pursuit: Learning inverse problems without ground-truth data
• Optimization methods for MR image reconstruction (long version)

Cited by in corpus (18)

• Self-Supervised Learning of Physics-Guided Reconstruction Neural Networks without Fully-Sampled Reference Data
• Adaptive Diffusion Priors for Accelerated MRI Reconstruction
• RARE: Image Reconstruction using Deep Priors Learned without Ground Truth
• Multi-Mask Self-Supervised Learning for Physics-Guided Neural Networks in Highly Accelerated MRI
• PUERT: Probabilistic Under-sampling and Explicable Reconstruction Network for CS-MRI
• 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
• CoIL: Coordinate-based Internal Learning for Imaging Inverse Problems
• Deep Learning for Biomedical Image Reconstruction: A Survey
• A Densely Interconnected Network for Deep Learning Accelerated MRI
• Learning Task-Specific Strategies for Accelerated MRI
• A Tutorial on MRI Reconstruction: From Modern Methods to Clinical Implications
• On Instabilities of Conventional Multi-Coil MRI Reconstruction to Small Adverserial Perturbations
• Improved Supervised Training of Physics-Guided Deep Learning Image Reconstruction with Multi-Masking
• Noise2Inpaint: Learning Referenceless Denoising by Inpainting Unrolling
• Deep Multi-view Depth Estimation with Predicted Uncertainty
• Overcoming Distribution Shifts in Plug-and-Play Methods with Test-Time Training