Score-based Self-supervised MRI Denoising

2025-05-08

AI Summary

MRI acceleration and low-field acquisition are highly susceptible to noise, degrading image quality and diagnostic reliability. Supervised denoising methods rely on high-SNR ground-truth labels that are difficult to obtain in practice, while existing self-supervised approaches often oversmooth fine details and suffer from limited performance. This paper proposes C2S, a label-free self-supervised denoising framework that introduces Generalized Denoising Score Matching (GDSM) loss—the first of its kind—to model the conditional expectation of high-SNR images directly from single-noisy acquisitions. By integrating noise-level reparameterization and a detail-refinement module, C2S jointly optimizes denoising strength and structural fidelity. Evaluated on M4Raw and fastMRI, C2S achieves state-of-the-art performance among self-supervised methods, matching supervised baselines while significantly improving texture preservation and cross-contrast generalizability.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Magnetic resonance imaging (MRI) is a powerful noninvasive diagnostic imaging tool that provides unparalleled soft tissue contrast and anatomical detail. Noise contamination, especially in accelerated and/or low-field acquisitions, can significantly degrade image quality and diagnostic accuracy. Supervised learning based denoising approaches have achieved impressive performance but require high signal-to-noise ratio (SNR) labels, which are often unavailable. Self-supervised learning holds promise to address the label scarcity issue, but existing self-supervised denoising methods tend to oversmooth fine spatial features and often yield inferior performance than supervised methods. We introduce Corruption2Self (C2S), a novel score-based self-supervised framework for MRI denoising. At the core of C2S is a generalized denoising score matching (GDSM) loss, which extends denoising score matching to work directly with noisy observations by modeling the conditional expectation of higher-SNR images given further corrupted observations. This allows the model to effectively learn denoising across multiple noise levels directly from noisy data. Additionally, we incorporate a reparameterization of noise levels to stabilize training and enhance convergence, and introduce a detail refinement extension to balance noise reduction with the preservation of fine spatial features. Moreover, C2S can be extended to multi-contrast denoising by leveraging complementary information across different MRI contrasts. We demonstrate that our method achieves state-of-the-art performance among self-supervised methods and competitive results compared to supervised counterparts across varying noise conditions and MRI contrasts on the M4Raw and fastMRI dataset.

Problem

• Self-supervised MRI denoising without high-SNR labels
• Preserving fine spatial features during noise reduction
• Extending denoising to multi-contrast MRI using complementary information

Innovation

• Score-based self-supervised MRI denoising framework
• Generalized denoising score matching loss function
• Multi-contrast denoising with complementary information