Abstract:In recent years, deep learning has attracted increasing at- tention in the field of Cardiac MRI (CMR) reconstruction due to its superior performance over traditional methods, particularly in handling higher acceleration factors, highlighting its potential for real-world clini- cal applications. However, current deep learning methods remain limited in generalizability. CMR scans exhibit wide variability in image contrast, sampling patterns, scanner vendors, anatomical structures, and disease types. Most existing models are designed to handle only a single or nar- row subset of these variations, leading to performance degradation when faced with distribution shifts. The…
Abstract:In recent years, deep learning has attracted increasing at- tention in the field of Cardiac MRI (CMR) reconstruction due to its superior performance over traditional methods, particularly in handling higher acceleration factors, highlighting its potential for real-world clini- cal applications. However, current deep learning methods remain limited in generalizability. CMR scans exhibit wide variability in image contrast, sampling patterns, scanner vendors, anatomical structures, and disease types. Most existing models are designed to handle only a single or nar- row subset of these variations, leading to performance degradation when faced with distribution shifts. Therefore, it is beneficial to develop a unified model capable of generalizing across diverse CMR scenarios. To this end, we propose CRUNet-MR-Univ, a foundation model that lever- ages spatio-temporal correlations and prompt-based priors to effectively handle the full diversity of CMR scans. Our approach consistently out- performs baseline methods across a wide range of settings, highlighting its effectiveness and promise.
| Comments: | STACOM 2025 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2601.04428 [cs.CV] |
| (or arXiv:2601.04428v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2601.04428 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Donghang Lyu [view email] [v1] Wed, 7 Jan 2026 22:23:56 UTC (2,377 KB)