Conditional Denoising Diffusion Model-Based Robust MR Image Reconstruction from Highly Undersampled Data

2025-10-07

AI Summary

In accelerated MRI reconstruction, severe aliasing artifacts arise from undersampled k-space acquisitions; existing diffusion models often treat data consistency as a post-processing step or rely on unsupervised objectives. This paper proposes a physics-informed conditional denoising diffusion model that explicitly incorporates the MRI forward measurement model into every reverse-diffusion step, enabling iterative data-consistency correction. By integrating paired supervised training with joint optimization of the generative prior and physical constraints, the method simultaneously enhances representational capacity and physical interpretability. Evaluated on the fastMRI dataset, our approach achieves state-of-the-art performance across all major metrics—SSIM, PSNR, and LPIPS—with particularly notable improvements in structural fidelity and perceptual quality. The framework establishes a new paradigm for robust, high-fidelity accelerated MRI reconstruction.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

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
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Magnetic Resonance Imaging (MRI) is a critical tool in modern medical diagnostics, yet its prolonged acquisition time remains a critical limitation, especially in time-sensitive clinical scenarios. While undersampling strategies can accelerate image acquisition, they often result in image artifacts and degraded quality. Recent diffusion models have shown promise for reconstructing high-fidelity images from undersampled data by learning powerful image priors; however, most existing approaches either (i) rely on unsupervised score functions without paired supervision or (ii) apply data consistency only as a post-processing step. In this work, we introduce a conditional denoising diffusion framework with iterative data-consistency correction, which differs from prior methods by embedding the measurement model directly into every reverse diffusion step and training the model on paired undersampled-ground truth data. This hybrid design bridges generative flexibility with explicit enforcement of MRI physics. Experiments on the fastMRI dataset demonstrate that our framework consistently outperforms recent state-of-the-art deep learning and diffusion-based methods in SSIM, PSNR, and LPIPS, with LPIPS capturing perceptual improvements more faithfully. These results demonstrate that integrating conditional supervision with iterative consistency updates yields substantial improvements in both pixel-level fidelity and perceptual realism, establishing a principled and practical advance toward robust, accelerated MRI reconstruction.

Problem

• Reconstructing high-fidelity MR images from undersampled data
• Addressing image artifacts from accelerated MRI acquisition
• Integrating data consistency directly into diffusion reconstruction steps

Innovation

• Conditional denoising diffusion with iterative data-consistency correction
• Embedding measurement model into every reverse diffusion step
• Training on paired undersampled-ground truth data