The R2D2 Deep Neural Network Series for Scalable Non-Cartesian Magnetic Resonance Imaging

2025-03-12

AI Summary

Embedding non-uniform FFT (NUFFT) in deep neural networks (DNNs) for non-Cartesian MRI reconstruction leads to impractical training, slow convergence, and poor scalability. Method: We propose R2D2, a novel deep learning paradigm that employs cascaded residual modules to iteratively estimate k-space data residuals, integrating matching pursuit principles into a learnable architecture. By avoiding direct NUFFT embedding, R2D2 reconciles the scalability of plug-and-play methods with the data consistency of unrolled networks. The framework jointly incorporates multi-coil parallel imaging, non-Cartesian k-space modeling, and supervised per-module training. Contribution/Results: Experiments under radial undersampling demonstrate that R2D2 achieves superior image quality and faster reconstruction than R2D2-Net and decomposition-based diffusion samplers—even with only a few modules—on both simulated and in vivo datasets.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

We introduce the R2D2 Deep Neural Network (DNN) series paradigm for fast and scalable image reconstruction from highly-accelerated non-Cartesian k-space acquisitions in Magnetic Resonance Imaging (MRI). While unrolled DNN architectures provide a robust image formation approach via data-consistency layers, embedding non-uniform fast Fourier transform operators in a DNN can become impractical to train at large scale, e.g in 2D MRI with a large number of coils, or for higher-dimensional imaging. Plug-and-play approaches that alternate a learned denoiser blind to the measurement setting with a data-consistency step are not affected by this limitation but their highly iterative nature implies slow reconstruction. To address this scalability challenge, we leverage the R2D2 paradigm that was recently introduced to enable ultra-fast reconstruction for large-scale Fourier imaging in radio astronomy. R2D2's reconstruction is formed as a series of residual images iteratively estimated as outputs of DNN modules taking the previous iteration's data residual as input. The method can be interpreted as a learned version of the Matching Pursuit algorithm. A series of R2D2 DNN modules were sequentially trained in a supervised manner on the fastMRI dataset and validated for 2D multi-coil MRI in simulation and on real data, targeting highly under-sampled radial k-space sampling. Results suggest that a series with only few DNNs achieves superior reconstruction quality over its unrolled incarnation R2D2-Net (whose training is also much less scalable), and over the state-of-the-art diffusion-based"Decomposed Diffusion Sampler"approach (also characterised by a slower reconstruction process).

Problem

• Scalable image reconstruction in non-Cartesian MRI
• Overcoming training challenges in large-scale DNNs
• Fast reconstruction for highly-accelerated k-space acquisitions

Innovation

• R2D2 DNN series for scalable MRI reconstruction
• Iterative residual estimation using DNN modules
• Supervised training on fastMRI dataset for validation