Measurement Score-Based MRI Reconstruction with Automatic Coil Sensitivity Estimation

2025-09-22

AI Summary

Conventional compressed sensing parallel MRI reconstruction relies on pre-calibrated coil sensitivity maps (CSMs) and clean ground-truth images—both impractical in clinical settings. Method: We propose a diffusion-based framework that jointly models CSM estimation and measurement-fraction self-supervised learning directly in k-space, eliminating the need for CSM calibration or fully-sampled training data. Our approach enables end-to-end, physics-informed estimation of CSMs and posterior score optimization within a diffusion architecture. By integrating diffusion priors with the physical forward model, we employ stochastic sampling and joint posterior optimization over undersampled measurements. Results: Evaluated on the fastMRI multi-coil brain dataset, our method achieves reconstruction performance comparable to supervised diffusion models and significantly outperforms existing unsupervised approaches, demonstrating strong robustness under high acceleration factors and high clinical applicability.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search

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: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Security and Privacy: Large-scale security measurements

Abstract

Diffusion-based inverse problem solvers (DIS) have recently shown outstanding performance in compressed-sensing parallel MRI reconstruction by combining diffusion priors with physical measurement models. However, they typically rely on pre-calibrated coil sensitivity maps (CSMs) and ground truth images, making them often impractical: CSMs are difficult to estimate accurately under heavy undersampling and ground-truth images are often unavailable. We propose Calibration-free Measurement Score-based diffusion Model (C-MSM), a new method that eliminates these dependencies by jointly performing automatic CSM estimation and self-supervised learning of measurement scores directly from k-space data. C-MSM reconstructs images by approximating the full posterior distribution through stochastic sampling over partial measurement posterior scores, while simultaneously estimating CSMs. Experiments on the multi-coil brain fastMRI dataset show that C-MSM achieves reconstruction performance close to DIS with clean diffusion priors -- even without access to clean training data and pre-calibrated CSMs.

Problem

• Eliminates dependency on pre-calibrated coil sensitivity maps
• Enables MRI reconstruction without ground truth training images
• Performs joint CSM estimation and self-supervised k-space learning

Innovation

• Joint automatic CSM estimation and self-supervised measurement score learning
• Stochastic sampling over partial measurement posterior scores
• Eliminates need for clean training data and pre-calibrated CSMs