CMRVision: A Foundation Model for Cardiac MR Image Analysis

• URL: http://arxiv.org/abs/2609.01308v1
• Date: Tue, 01 Sep 2026 14:34:30 GMT
• Title: CMRVision: A Foundation Model for Cardiac MR Image Analysis
• Abstract summary: We introduce CMRVision, a CMR-specific foundation model trained using DINOv3-style self-supervised learning.<n>CMRVision is evaluated on two downstream tasks: multi-task segmentation across cine, late gadolinium enhancement (LGE) and mapping sequences.<n>Our experiments show that CMR-specific pretraining, smaller patch sizes, and patch-level objectives consistently improve downstream performance.
• Score: 21.180056139848528
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Cardiac magnetic resonance (CMR) imaging provides complementary information on cardiac anatomy, function, and tissue characterization across multiple sequences and views. In this work, we investigate foundation model pretraining for 2D CMR and introduce CMRVision, a CMR-specific foundation model trained using DINOv3-style self-supervised learning on a multi-center, multi-sequence cohort of 36 million CMR images. We systematically evaluate architectural and training design choices for domain-specific pretraining. CMRVision is evaluated on two downstream tasks: multi-task segmentation across cine, late gadolinium enhancement (LGE), and mapping sequences, and cine view classification. Our experiments show that CMR-specific pretraining, smaller patch sizes, and patch-level objectives consistently improve downstream performance. Across a multi-task segmentation benchmark, CMRVision achieved the strongest overall performance, outperforming prior natural-image (NI), medical-image, supervised, and CMR foundation model baselines. Improvements were modest but consistent across structures and sequences, with Dice scores ranging from 0.940-0.967 for LV and 0.855-0.905 for myocardium, and reaching 0.929 for RV, 0.920 for LA, and 0.931 for RA. The largest gains were observed for myocardium segmentation in LGE and mapping images. In a zero-shot segmentation task on unseen LGE long-axis views, the model achieved an average Dice score of 0.692, demonstrating cross-view generalization. For cine view classification, CMRVision achieved the highest average accuracy (0.906), compared to prior methods reported in the literature. These results highlight the potential of CMRVision to support robust and generalizable cardiac MRI analysis across multiple sequences and views.

Related papers

• MR-JEPA: A General Purpose Video Foundation Model for Cardiac MRI [21.245485191371703]
  We present MR-JEPA, a self-supervised video foundation model for CMR.<n>Unlike prior CMR video models limited to cine data, MR-JEPA is pretrained on sequential multi-sequence data.
  arXiv  Detail & Related papers  (2026-08-31T15:36:24Z)
• Foundation Models Adaptation for Multi-View Multi-modal Cardiac MRI Segmentation and Direct Ejection Fraction Estimation [0.3957682607484472]
  Foundation models have shown strong transferability in cardiac MRI (CMR), but their effectiveness for heterogeneous multi-view and multi-sequence analysis remains unclear.<n>We explore the effectiveness of fine-tuning and combining different CMR foundation models for the Universal Multi-Sequence, Multi-Center and Multi-View CMR (CMR-Multi) Challenge.
  arXiv  Detail & Related papers  (2026-08-07T14:48:51Z)
• ConvFormer3D-TAP: Phase/Uncertainty-Aware Front-End Fusion for Cine CMR View Classification Pipelines [0.879504058268139]
  ConvFormer3D-TAP integrates 3D convolutional tokenization with multiscale self-attention to enhance robustness across cardiac phases and ambiguous temporal segments.<n>On a cohort of 150,974 clinically acquired cine sequences, ConvFormer3D- achieved 96% validation accuracy with per-class F1-scores.<n>These results support ConvFormer3DTAP as a scalable front-end for view routing, filtering and quality control in end-to-end cMRI.
  arXiv  Detail & Related papers  (2026-04-13T12:29:16Z)
• Enabling Ultra-Fast Cardiovascular Imaging Across Heterogeneous Clinical Environments with a Generalist Foundation Model and Multimodal Database [64.65360708629485]
  MMCMR-427K is the largest and most comprehensive multimodal cardiovascular magnetic resonance k-space database.<n> CardioMM is a reconstruction foundation model capable of adapting to heterogeneous fast CMR imaging scenarios.<n> CardioMM unifies semantic contextual understanding with physics-informed data consistency to deliver robust reconstructions.
  arXiv  Detail & Related papers  (2025-12-25T12:47:50Z)
• LGE-Guided Cross-Modality Contrastive Learning for Gadolinium-Free Cardiomyopathy Screening in Cine CMR [51.11296719862485]
  We propose a Contrastive Learning and Cross-Modal alignment framework for gadolinium-free cardiomyopathy screening using cine CMR sequences.<n>By aligning the latent spaces of cine CMR and Late Gadolinium Enhancement (LGE) sequences, our model encodes fibrosis-specific pathology into cine CMR embeddings.
  arXiv  Detail & Related papers  (2025-08-23T07:21:23Z)
• A versatile foundation model for cine cardiac magnetic resonance image analysis tasks [6.488550274514015]
  We present a versatile foundation model that can perform a range of clinically-relevant image analysis tasks.<n>A multi-view convolution-transformer masked autoencoder, named as CineMA, was trained on 15 million cine images from 74,916 subjects.
  arXiv  Detail & Related papers  (2025-05-31T19:12:34Z)
• TotalSegmentator MRI: Robust Sequence-independent Segmentation of Multiple Anatomic Structures in MRI [59.86827659781022]
  A nnU-Net model (TotalSegmentator) was trained on MRI and segment 80atomic structures.<n>Dice scores were calculated between the predicted segmentations and expert reference standard segmentations to evaluate model performance.<n>Open-source, easy-to-use model allows for automatic, robust segmentation of 80 structures.
  arXiv  Detail & Related papers  (2024-05-29T20:15:54Z)
• Multi-view Hybrid Graph Convolutional Network for Volume-to-mesh Reconstruction in Cardiovascular MRI [43.47826598981827]
  We introduce HybridVNet, a novel architecture for direct image-to-mesh extraction.<n>We show it can efficiently handle surface and volumetric meshes by encoding them as graph structures.<n>Our model combines traditional convolutional networks with variational graph generative models, deep supervision and mesh-specific regularisation.
  arXiv  Detail & Related papers  (2023-11-22T21:51:29Z)
• CMRxRecon: An open cardiac MRI dataset for the competition of accelerated image reconstruction [62.61209705638161]
  There has been growing interest in deep learning-based CMR imaging algorithms. Deep learning methods require large training datasets. This dataset includes multi-contrast, multi-view, multi-slice and multi-coil CMR imaging data from 300 subjects.
  arXiv  Detail & Related papers  (2023-09-19T15:14:42Z)
• LVM-Med: Learning Large-Scale Self-Supervised Vision Models for Medical Imaging via Second-order Graph Matching [59.01894976615714]
  We introduce LVM-Med, the first family of deep networks trained on large-scale medical datasets. We have collected approximately 1.3 million medical images from 55 publicly available datasets. LVM-Med empirically outperforms a number of state-of-the-art supervised, self-supervised, and foundation models.
  arXiv  Detail & Related papers  (2023-06-20T22:21:34Z)
• Deep Learning Framework with Multi-Head Dilated Encoders for Enhanced Segmentation of Cervical Cancer on Multiparametric Magnetic Resonance Imaging [0.6597195879147557]
  T2-weighted magnetic resonance imaging (MRI) and diffusion-weighted imaging (DWI) are essential components for cervical cancer diagnosis. We propose a novel multi-head framework that uses dilated convolutions and shared residual connections for separate encoding of multiparametric MRI images.
  arXiv  Detail & Related papers  (2023-06-19T19:41:21Z)
• A Global Benchmark of Algorithms for Segmenting Late Gadolinium-Enhanced Cardiac Magnetic Resonance Imaging [90.29017019187282]
  " 2018 Left Atrium Challenge" using 154 3D LGE-MRIs, currently the world's largest cardiac LGE-MRI dataset. Analyse of the submitted algorithms using technical and biological metrics was performed. Results show the top method achieved a dice score of 93.2% and a mean surface to a surface distance of 0.7 mm.
  arXiv  Detail & Related papers  (2020-04-26T08:49:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.