An unsupervised method for MRI recovery: Deep image prior with structured sparsity

2025-01-02

AI Summary

To address the challenge of lacking fully-sampled k-space data and paired ground-truth labels in dynamic MRI reconstruction, this paper proposes an unsupervised blind reconstruction method. The approach integrates structured group sparsity regularization into the Deep Image Prior (DIP) framework for the first time, imposing group sparsity constraints on frame-correlated latent code vectors to adaptively learn a low-dimensional temporal manifold. Critically, it requires no labeled data or paired training samples. Evaluated on single-shot late gadolinium enhancement (LGE) dynamic MRI, the method achieves high-fidelity reconstructions under severe undersampling. Quantitative results demonstrate a 32% reduction in normalized mean squared error (NMSE) and a 0.08 increase in structural similarity index (SSIM) compared to conventional compressed sensing and standard DIP. Moreover, clinical expert blind assessment ranks our method highest, confirming its superior accuracy and clinical utility for undersampled dynamic MRI reconstruction.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Multi-modal Vision
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Objective: To propose and validate an unsupervised MRI reconstruction method that does not require fully sampled k-space data. Materials and Methods: The proposed method, deep image prior with structured sparsity (DISCUS), extends the deep image prior (DIP) by introducing group sparsity to frame-specific code vectors, enabling the discovery of a low-dimensional manifold for capturing temporal variations. discus was validated using four studies: (I) simulation of a dynamic Shepp-Logan phantom to demonstrate its manifold discovery capabilities, (II) comparison with compressed sensing and DIP-based methods using simulated single-shot late gadolinium enhancement (LGE) image series from six distinct digital cardiac phantoms in terms of normalized mean square error (NMSE) and structural similarity index measure (SSIM), (III) evaluation on retrospectively undersampled single-shot LGE data from eight patients, and (IV) evaluation on prospectively undersampled single-shot LGE data from eight patients, assessed via blind scoring from two expert readers. Results: DISCUS outperformed competing methods, demonstrating superior reconstruction quality in terms of NMSE and SSIM (Studies I--III) and expert reader scoring (Study IV). Discussion: An unsupervised image reconstruction method is presented and validated on simulated and measured data. These developments can benefit applications where acquiring fully sampled data is challenging.

Problem

• MRI image reconstruction
• incomplete data
• autonomous learning

Innovation

• Deep Image Prior
• Organized Sparsity
• Self-learning MRI Reconstruction