Score-based Generative Priors Guided Model-driven Network for MRI Reconstruction
Not on the map. Why?
Abstract (arXiv)
Score matching with Langevin dynamics (SMLD) method has been successfully applied to accelerated MRI. However, the hyperparameters in the sampling process require subtle tuning, otherwise the results can be severely corrupted by hallucination artifacts, especially with out-of-distribution test data. To address the limitations, we proposed a novel workflow where naive SMLD samples serve as additional priors to guide model-driven network training. First, we adopted a pretrained score network to generate samples as preliminary guidance images (PGI), obviating the need for network retraining, parameter tuning and in-distribution test data. Although PGIs are corrupted by hallucination artifacts, we believe they can provide extra information through effective denoising steps to facilitate reconstruction. Therefore, we designed a denoising module (DM) in the second step to coarsely eliminate artifacts and noises from PGIs. The features are extracted from a score-based information extractor (SIE) and a cross-domain information extractor (CIE), which directly map to the noise patterns. Third, we designed a model-driven network guided by denoised PGIs (DGIs) to further recover fine details. DGIs are densely connected with intermediate reconstructions in each cascade to enrich the information and are periodically updated to provide more accurate guidance. Our experiments on different datasets reveal that despite the low average quality of PGIs, the proposed workflow can effectively extract valuable information to guide the network training, even with severely reduced training data and sampling steps. Our method outperforms other cutting-edge techniques by effectively mitigating hallucination artifacts, yielding robust and high-quality reconstruction results.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 52 of its 132 references that are in the map. First posted on arXiv 5 May 2024.
Builds on
52 of 132 references.
References · 52
- 2020Denoising Diffusion Probabilistic Modelscited by 21,076
- 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
- 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,101
- 2020Improved Techniques for Training Score-Based Generative Modelscited by 1,043
- 2018fastMRI: An Open Dataset and Benchmarks for Accelerated MRIcited by 470
- 2016RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentationcited by 465
- 2016Deep Convolutional Neural Network for Inverse Problems in Imagingcited by 433
- 2017MoDL: Model Based Deep Learning Architecture for Inverse Problemscited by 387
- 2017Learning a Variational Network for Reconstruction of Accelerated MRI Datacited by 342
- 2017Compressed Sensing using Generative Modelscited by 339
- 2021Score-based diffusion models for accelerated MRIcited by 311
- 2021Robust Compressed Sensing MRI with Deep Generative Priorscited by 286
38 more
- 2022Cold Diffusion: Inverting Arbitrary Image Transforms Without Noisecited by 276
- 2022Diffusion Models for Medical Image Analysis: A Comprehensive Surveycited by 273
- 2016A Kernelized Stein Discrepancy for Goodness-of-fit Tests and Model Evaluationcited by 236
- 2017A Deep Cascade of Convolutional Neural Networks for Dynamic MR Image Reconstructioncited by 232
- 2020A Review on Generative Adversarial Networks: Algorithms, Theory, and Applicationscited by 209
- 2023A Survey on Self-supervised Learning: Algorithms, Applications, and Future Trendscited by 207
- 2020End-to-End Variational Networks for Accelerated MRI Reconstructioncited by 202
- 2022Adaptive Diffusion Priors for Accelerated MRI Reconstructioncited by 118
- 2017Compressed Sensing MRI Reconstruction using a Generative Adversarial Network with a Cyclic Losscited by 82
- 2022Towards performant and reliable undersampled MR reconstruction via diffusion model samplingcited by 60
- 2018Deep Residual Learning for Accelerated MRI using Magnitude and Phase Networkscited by 53
- 2022ReconFormer: Accelerated MRI Reconstruction Using Recurrent Transformercited by 52
- 2019Neural Empirical Bayescited by 46
- 2022High-Frequency Space Diffusion Models for Accelerated MRIcited by 42
- 2021Memory-Augmented Deep Unfolding Network for Compressive Sensingcited by 38
- 2023Self-Supervised MRI Reconstruction with Unrolled Diffusion Modelscited by 38
- 2021Recurrent Variational Network: A Deep Learning Inverse Problem Solver applied to the task of Accelerated MRI Reconstructioncited by 34
- 2019VS-Net: Variable splitting network for accelerated parallel MRI reconstructioncited by 31
- 2019Deep Generalization of Structured Low-Rank Algorithms (Deep-SLR)cited by 30
- 2022HUMUS-Net: Hybrid unrolled multi-scale network architecture for accelerated MRI reconstructioncited by 30
- 2021One-dimensional Deep Low-rank and Sparse Network for Accelerated MRIcited by 27
- 2019Learning Energy-Based Models in High-Dimensional Spaces with Multi-scale Denoising Score Matchingcited by 26
- 2020Deep learning for fast MR imaging: a review for learning reconstruction from incomplete k-space datacited by 24
- 2023IMJENSE: Scan-specific Implicit Representation for Joint Coil Sensitivity and Image Estimation in Parallel MRIcited by 23
- 2023Diffusion Modeling with Domain-conditioned Prior Guidance for Accelerated MRI and qMRI Reconstructioncited by 22
- 2021Deep J-Sense: Accelerated MRI Reconstruction via Unrolled Alternating Optimizationcited by 19
- 2021Accelerated MRI Reconstruction with Separable and Enhanced Low-Rank Hankel Regularizationcited by 18
- 2023SMRD: SURE-based Robust MRI Reconstruction with Diffusion Modelscited by 16
- 2019Pyramid Convolutional RNN for MRI Image Reconstructioncited by 14
- 2020Homotopic Gradients of Generative Density Priors for MR Image Reconstructioncited by 14
- 2023Correlated and Multi-frequency Diffusion Modeling for Highly Under-sampled MRI Reconstructioncited by 14
- 2024Progressive Divide-and-Conquer via Subsampling Decomposition for Accelerated MRIcited by 13
- 2023CDiffMR: Can We Replace the Gaussian Noise with K-Space Undersampling for Fast MRI?cited by 12
- 2022Self-supervised Deep Unrolled Reconstruction Using Regularization by Denoisingcited by 8
- 2023Global k-Space Interpolation for Dynamic MRI Reconstruction using Masked Image Modelingcited by 8
- 2023vSHARP: variable Splitting Half-quadratic Admm algorithm for Reconstruction of inverse-Problemscited by 8
- 2023Fast Controllable Diffusion Models for Undersampled MRI Reconstructioncited by 8
- 2019Efficient Structurally-Strengthened Generative Adversarial Network for MRI Reconstructioncited by 5
50 most cited of 52 All 52 (JSON)
Built on by
Citing papers per year: 2024: 1; 2025: 1; 2026: 1.
3 papers.