DeepMpMRI: Tensor-decomposition Regularized Learning for Fast and High-Fidelity Multi-Parametric Microstructural MR Imaging

2024-05-06

arXiv.org

AI Summary

To address the low efficiency and poor fidelity in joint estimation of multiple microstructural model parameters from sparsely sampled q-space data in multi-parametric microstructure MRI, this paper proposes a deep learning framework incorporating tensor decomposition regularization. Our key contributions are: (1) a novel tensor decomposition regularization scheme that explicitly encodes structural correlations across parameter maps; (2) a Nesterov-accelerated adaptive learning algorithm for dynamic regularization strength optimization; and (3) a scalable, architecture-flexible end-to-end reconstruction network. Quantitative evaluation demonstrates consistent superiority over five state-of-the-art methods, with significant improvements in accuracy metrics (e.g., PSNR, SSIM) and faithful preservation of fine microstructural details in qualitative results. Compared to full 270-direction acquisitions, our method achieves acceleration factors of 4.5×–22.5× while maintaining high reconstruction fidelity.

Technology Category

• Machine Learning: Matrix & Tensor Methods
• Computer Vision: Learning & Optimization for CV
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Deep learning has emerged as a promising approach for learning the nonlinear mapping between diffusion-weighted MR images and tissue parameters, which enables automatic and deep understanding of the brain microstructures. However, the efficiency and accuracy in the multi-parametric estimations are still limited since previous studies tend to estimate multi-parametric maps with dense sampling and isolated signal modeling. This paper proposes DeepMpMRI, a unified framework for fast and high-fidelity multi-parametric estimation from various diffusion models using sparsely sampled q-space data. DeepMpMRI is equipped with a newly designed tensor-decomposition-based regularizer to effectively capture fine details by exploiting the correlation across parameters. In addition, we introduce a Nesterov-based adaptive learning algorithm that optimizes the regularization parameter dynamically to enhance the performance. DeepMpMRI is an extendable framework capable of incorporating flexible network architecture. Experimental results demonstrate the superiority of our approach over 5 state-of-the-art methods in simultaneously estimating multi-parametric maps for various diffusion models with fine-grained details both quantitatively and qualitatively, achieving 4.5 - 22.5$ imes$ acceleration compared to the dense sampling of a total of 270 diffusion gradients.

Problem

• Estimating multiple microstructural parameters from sparse MRI data
• Improving efficiency and accuracy in multi-model parameter mapping
• Enhancing detail capture via tensor-decomposition regularization

Innovation

• Tensor-decomposition regularizer captures fine details
• Nesterov-based adaptive learning optimizes parameters dynamically
• Extendable framework for flexible network architecture integration