Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

MRI-Based Pressure Gradient Mapping in Patient-Specific Models of Coarctation of the Aorta.

medRxiv : the preprint server for health sciences·2026
Same author

Diffusion-weighted Imaging of the Liver: Primed for New Business.

Radiology·2026
Same author

Personalized biventricular mechanics and sensitivity to model morphology.

bioRxiv : the preprint server for biology·2025
Same author

Toward Modality- and Sampling-Universal Learning Strategies for Accelerating Cardiovascular Imaging: Summary of the CMRxRecon2024 Challenge.

IEEE transactions on medical imaging·2025
Same author

The Effect of Voxel Volume and Voxel Shape on Cardiac Diffusion Tensor Imaging Metrics.

Magnetic resonance in medicine·2025
Same author

Regional heterogeneity in left atrial stiffness impacts passive deformation in a cohort of patient-specific models.

PLoS computational biology·2025

Related Experiment Video

Updated: Aug 9, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.9K

Myocardial Segmentation of Tagged Magnetic Resonance Images with Transfer Learning Using Generative Cine-To-Tagged

Arnaud P Dhaene1,2, Michael Loecher1,3, Alexander J Wilson1,3

  • 1Department of Radiology, Stanford University, Stanford, CA 94305, USA.

Bioengineering (Basel, Switzerland)
|February 25, 2023
PubMed
Summary

Deep learning models can now segment cardiac tagged MRI images, automating a key step for myocardial strain analysis. This advancement streamlines workflows by enabling the use of large public datasets for training.

Keywords:
MRI taggingcardiac MRIdataset augmentationdeep learningdomain transformationmyocardial segmentation

More Related Videos

Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

9.1K
Analyzing Mitochondrial Morphology Through Simulation Supervised Learning
12:06

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Published on: March 3, 2023

4.1K

Related Experiment Videos

Last Updated: Aug 9, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

2.9K
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
10:25

Deep Learning-Based Segmentation of Cryo-Electron Tomograms

Published on: November 11, 2022

9.1K
Analyzing Mitochondrial Morphology Through Simulation Supervised Learning
12:06

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Published on: March 3, 2023

4.1K

Area of Science:

  • Medical imaging
  • Artificial intelligence
  • Cardiovascular research

Background:

  • Deep learning (DL) segmentation in cardiac MRI can improve radiology workflow efficiency, especially for myocardial strain measurement.
  • Current DL motion tracking models accelerate displacement field measurement and strain estimation.
  • Manual input for reference point selection in myocardial segmentation remains a bottleneck, particularly for tagged cardiac magnetic resonance (CMR) images.

Purpose of the Study:

  • To develop and compare two novel DL models for myocardium segmentation from tagged CMR images.
  • To adapt large public cine CMR datasets for training segmentation models for tagged images.
  • To evaluate the effectiveness of different cine-to-tagged image transformation methods.

Main Methods:

  • Developed and compared nnU-net and Segmentation ResNet VAE models for tagged CMR segmentation.
  • Implemented physics-driven and generative adversarial network (GAN) style transfer models for cine-to-tagged image transformation.
  • Utilized pretraining strategies with transformed cine images to leverage public datasets.

Main Results:

  • Pretrained models demonstrated superior performance (+2.8 Dice) and faster convergence (6×) compared to models trained from scratch.
  • The best results were achieved using a generative model for cine-to-tagged image transformation, preserving structure.
  • The developed segmentation network achieved a Dice coefficient of 0.828 and a 95th percentile Hausdorff distance of 4.745 mm.

Conclusions:

  • Novel DL models successfully segment myocardium from tagged CMR images, achieving performance comparable to cine image segmentation.
  • Pretraining on transformed cine images significantly enhances model performance and training efficiency.
  • This work enables automated reference point initialization for myocardial strain analysis using tagged CMR.