Unsupervised Motion-Compensated Decomposition for Cardiac MRI Reconstruction via Neural Representation

2025-11-14

AI Summary

To address the challenges of low reconstruction quality, reliance on ground-truth labels, and respiratory motion artifacts in highly undersampled k-t space dynamic cardiac MR imaging, this paper proposes an unsupervised MoCo-INR framework. It synergistically integrates implicit neural representations (INRs) with an explicit motion compensation module to enable accurate, label-free cardiac motion decomposition and continuous spatiotemporal modeling. The method employs a customized network architecture optimized end-to-end directly in the k-t domain, eliminating dependence on scarce paired ground-truth data inherent to supervised learning. Evaluated on both simulated and real free-breathing datasets, MoCo-INR achieves up to 20× acceleration while preserving fine anatomical details, demonstrating rapid convergence and high fidelity. Its computational efficiency and reconstruction accuracy indicate strong potential for clinical real-time dynamic cardiac MR imaging.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Calibration & Uncertainty Quantification
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• 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
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Cardiac magnetic resonance (CMR) imaging is widely used to characterize cardiac morphology and function. To accelerate CMR imaging, various methods have been proposed to recover high-quality spatiotemporal CMR images from highly undersampled k-t space data. However, current CMR reconstruction techniques either fail to achieve satisfactory image quality or are restricted by the scarcity of ground truth data, leading to limited applicability in clinical scenarios. In this work, we proposed MoCo-INR, a new unsupervised method that integrates implicit neural representations (INR) with the conventional motion-compensated (MoCo) framework. Using explicit motion modeling and the continuous prior of INRs, MoCo-INR can produce accurate cardiac motion decomposition and high-quality CMR reconstruction. Furthermore, we introduce a new INR network architecture tailored to the CMR problem, which significantly stabilizes model optimization. Experiments on retrospective (simulated) datasets demonstrate the superiority of MoCo-INR over state-of-the-art methods, achieving fast convergence and fine-detailed reconstructions at ultra-high acceleration factors (e.g., 20x in VISTA sampling). Additionally, evaluations on prospective (real-acquired) free-breathing CMR scans highlight the clinical practicality of MoCo-INR for real-time imaging. Several ablation studies further confirm the effectiveness of the critical components of MoCo-INR.

Problem

• Reconstructing high-quality cardiac MRI from undersampled data
• Overcoming limitations of current methods without ground truth
• Achieving accurate motion decomposition for clinical practicality

Innovation

• Unsupervised motion-compensated decomposition via neural representation
• Implicit neural representations integrated with motion modeling
• Novel network architecture stabilizing cardiac MRI optimization