Non-rigid Motion Correction for MRI Reconstruction via Coarse-To-Fine Diffusion Models

2025-05-21

AI Summary

MRI’s prolonged acquisition time often induces non-rigid motion artifacts, severely degrading diagnostic quality in dynamic imaging. To address this, we propose a prior-free alternating optimization framework that jointly reconstructs images and corrects non-rigid k-space distortions. Our key innovation is a novel “coarse-to-fine” diffusion denoising strategy: first, a customized multi-scale diffusion model recovers low-frequency structural priors to robustly guide motion estimation; second, k-space domain deformation modeling—integrated with frequency-guided reconstruction—is optimized end-to-end within an alternating minimization framework. The method imposes no assumptions on sampling pattern, anatomical structure, or scanning protocol. Evaluated on 64× undersampled cardiac cine MRI and challenging simulated non-rigid deformations, it achieves substantial suppression of motion artifacts while preserving diagnostic-level image fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Non-convex Optimization
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Magnetic Resonance Imaging (MRI) is highly susceptible to motion artifacts due to the extended acquisition times required for k-space sampling. These artifacts can compromise diagnostic utility, particularly for dynamic imaging. We propose a novel alternating minimization framework that leverages a bespoke diffusion model to jointly reconstruct and correct non-rigid motion-corrupted k-space data. The diffusion model uses a coarse-to-fine denoising strategy to capture large overall motion and reconstruct the lower frequencies of the image first, providing a better inductive bias for motion estimation than that of standard diffusion models. We demonstrate the performance of our approach on both real-world cine cardiac MRI datasets and complex simulated rigid and non-rigid deformations, even when each motion state is undersampled by a factor of 64x. Additionally, our method is agnostic to sampling patterns, anatomical variations, and MRI scanning protocols, as long as some low frequency components are sampled during each motion state.

Problem

• Corrects non-rigid motion artifacts in MRI reconstruction
• Uses coarse-to-fine diffusion for motion estimation and denoising
• Works with undersampled data and diverse MRI protocols

Innovation

• Coarse-to-fine diffusion model for MRI motion correction
• Alternating minimization framework for joint reconstruction
• Agnostic to sampling patterns and anatomical variations