Are Natural-Domain Foundation Models Effective for Accelerated Cardiac MRI Reconstruction?

2026-04-24

AI Summary

This work investigates whether foundation vision models pretrained on natural domains—such as CLIP, DINOv2, and BiomedCLIP—can serve as effective priors to accelerate cardiac MRI reconstruction and evaluates their generalization capability across anatomical domains (e.g., heart → knee or brain). To this end, the authors propose an unrolled reconstruction framework that incorporates a frozen, pretrained vision encoder at each cascade stage to guide the optimization process, achieving high-quality reconstructions without fine-tuning. The study presents the first systematic comparison of transfer performance between natural-domain and medical-domain foundation models in accelerated MRI. The proposed method approaches state-of-the-art performance on in-distribution tasks and significantly enhances reconstruction robustness and generalization under high acceleration factors in cross-anatomical settings.

Technology Category

• Computer Vision: Other Foundations of Computer Vision
• Machine Learning: Deep Neural Architectures and Foundation Models
• Cognitive Modeling & Cognitive Systems: Other Foundations of Cognitive Modeling & Systems

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The emergence of large-scale pretrained foundation models has transformed computer vision, enabling strong performance across diverse downstream tasks. However, their potential for physics-based inverse problems, such as accelerated cardiac MRI reconstruction, remains largely underexplored. In this work, we investigate whether natural-domain foundation models can serve as effective image priors for accelerated cardiac MRI reconstruction, and compare the performance obtained against domain-specific counterparts such as BiomedCLIP. We propose an unrolled reconstruction framework that incorporates pretrained, frozen visual encoders, such as CLIP, DINOv2, and BiomedCLIP, within each cascade to guide the reconstruction process. Through extensive experiments, we show that while task-specific state-of-the-art reconstruction models such as E2E-VarNet achieve superior performance in standard in-distribution settings, foundation-model-based approaches remain competitive. More importantly, in challenging cross-domain scenarios, where models are trained on cardiac MRI and evaluated on anatomically distinct knee and brain datasets--foundation models exhibit improved robustness, particularly under high acceleration factors and limited low-frequency sampling. We further observe that natural-image-pretrained models, such as CLIP, learn highly transferable structural representations, while domain-specific pretraining (BiomedCLIP) provides modest additional gains in more ill-posed regimes. Overall, our results suggest that pretrained foundation models offer a promising source of transferable priors, enabling improved robustness and generalization in accelerated MRI reconstruction.

Problem

• accelerated MRI reconstruction
• foundation models
• cross-domain generalization
• inverse problems
• image priors

Innovation

• foundation models
• accelerated MRI reconstruction
• cross-domain generalization
• unrolled optimization
• transferable priors