Moner: Motion Correction in Undersampled Radial MRI with Unsupervised Neural Representation

2024-09-25

arXiv.org

AI Summary

In radial MRI, unpredictable rigid-body motion introduces severe artifacts in undersampled k-space reconstruction. Existing supervised motion correction (MoCo) methods rely on large-scale, high-quality labeled datasets, suffering from poor generalizability and high annotation costs. To address this, we propose the first fully unsupervised MoCo framework requiring no training data: it embeds a quasi-static rigid-body motion model into implicit neural representations (INRs), integrates a differentiable back-projection built upon the Fourier slice theorem, and employs a coarse-to-fine hash encoding to enhance geometric modeling fidelity. Our method jointly reconstructs artifact-free images and high-precision motion trajectories. Within-domain, it matches state-of-the-art supervised methods; in cross-domain scenarios, it significantly improves both motion estimation accuracy and image quality—achieving true zero-shot generalization.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Unsupervised & Self-Supervised Learning
• Intelligent Robots: Motion and Path Planning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Motion correction (MoCo) in radial MRI is a challenging problem due to the unpredictability of subject's motion. Current state-of-the-art (SOTA) MoCo algorithms often use extensive high-quality MR images to pre-train neural networks, obtaining excellent reconstructions. However, the need for large-scale datasets significantly increases costs and limits model generalization. In this work, we propose Moner, an unsupervised MoCo method that jointly solves artifact-free MR images and accurate motion from undersampled, rigid motion-corrupted k-space data, without requiring training data. Our core idea is to leverage the continuous prior of implicit neural representation (INR) to constrain this ill-posed inverse problem, enabling ideal solutions. Specifically, we incorporate a quasi-static motion model into the INR, granting its ability to correct subject's motion. To stabilize model optimization, we reformulate radial MRI as a back-projection problem using the Fourier-slice theorem. Additionally, we propose a novel coarse-to-fine hash encoding strategy, significantly enhancing MoCo accuracy. Experiments on multiple MRI datasets show our Moner achieves performance comparable to SOTA MoCo techniques on in-domain data, while demonstrating significant improvements on out-of-domain data. The code is available at: https://github.com/iwuqing/Moner

Problem

• Unsupervised motion correction in MRI
• Artifact-free image reconstruction
• Out-of-domain generalization improvement

Innovation

• Unsupervised motion correction method
• Implicit neural representation prior
• Coarse-to-fine hash encoding strategy