Self-Supervised Diffusion MRI Denoising via Iterative and Stable Refinement

2025-01-23

AI Summary

To address inaccurate microstructural modeling and fiber tractography caused by low signal-to-noise ratio (SNR) in diffusion MRI (dMRI), this paper proposes Di-Fusion—the first single-stage, fully self-supervised denoising framework that requires neither clean ground-truth labels nor explicit noise modeling. Methodologically, Di-Fusion integrates backward diffusion steps with an adaptive sampling strategy to construct an iterative refinement mechanism, and introduces a consistency-constrained self-supervised loss for stable training and controllable denoising. Experiments on both realistic and simulated dMRI datasets demonstrate that Di-Fusion significantly improves model fitting accuracy (R² increased by 12.7%), fractional anisotropy (FA) map quality, and fiber tractography completeness (fiber density cross-section, FDC, improved by 18.3%). Moreover, downstream tasks—including microstructural parameter estimation and white matter bundle reconstruction—achieve state-of-the-art performance.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Magnetic Resonance Imaging (MRI), including diffusion MRI (dMRI), serves as a ``microscope'' for anatomical structures and routinely mitigates the influence of low signal-to-noise ratio scans by compromising temporal or spatial resolution. However, these compromises fail to meet clinical demands for both efficiency and precision. Consequently, denoising is a vital preprocessing step, particularly for dMRI, where clean data is unavailable. In this paper, we introduce Di-Fusion, a fully self-supervised denoising method that leverages the latter diffusion steps and an adaptive sampling process. Unlike previous approaches, our single-stage framework achieves efficient and stable training without extra noise model training and offers adaptive and controllable results in the sampling process. Our thorough experiments on real and simulated data demonstrate that Di-Fusion achieves state-of-the-art performance in microstructure modeling, tractography tracking, and other downstream tasks.

Problem

• Diffusion MRI
• Noise Reduction
• Image Processing

Innovation

• Di-Fusion
• self-learning denoising
• diffusion MRI