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

Air Toxicity Surveillance across Thirteen Cities Using Rats.

Environmental science & technology·2025
Same author

HSPA12A stimulates "Smurf1-Hif1α-aerobic glycolysis" axis to promote proliferation of renal tubular epithelial cells after hypoxia/reoxygenation injury.

Cell stress & chaperones·2024
Same author

Data-driven machine learning models for predicting the maximum absorption and emission wavelengths of single benzene fluorophores.

Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy·2024
Same author

HSPA12A maintains aerobic glycolytic homeostasis and Histone3 lactylation in cardiomyocytes to attenuate myocardial ischemia/reperfusion injury.

JCI insight·2024
Same author

Microfluidic Brain-on-a-Chip: From Key Technology to System Integration and Application.

Small (Weinheim an der Bergstrasse, Germany)·2023
Same author

Comparison of Pharmacokinetic Similarity, Immunogenicity, and Safety of Ustekinumab and BAT2206 in Healthy Chinese Male Subjects in a Double-Blind, Randomized, Single-Dose, Parallel-Group Phase I Trial.

BioDrugs : clinical immunotherapeutics, biopharmaceuticals and gene therapy·2022

Related Experiment Video

Updated: Jul 27, 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

2.8K

A Double-Teacher Model Capable of Exploiting Isomorphic and Heterogeneous Discrepancy Information for Medical Image

Junguo Zou1, Zhaohe Wang2, Xiuquan Du2

  • 1School of Information Engineering, Chuzhou Polytechnic, Chuzhou 239000, China.

Diagnostics (Basel, Switzerland)
|June 10, 2023
PubMed
Summary

This study introduces a novel double-teacher deep learning framework for improved left atrial segmentation. The method leverages discrepancy information to enhance model performance without requiring fully 3D models, reducing memory and data demands.

Keywords:
heterogeneous discrepancy informationisomorphic discrepancy informationleft atrium segmentationsemi-supervised learning

More Related Videos

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
07:13

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities

Published on: October 27, 2023

1.2K
Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions
06:18

Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions

Published on: April 5, 2024

1.1K

Related Experiment Videos

Last Updated: Jul 27, 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

2.8K
Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
07:13

Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities

Published on: October 27, 2023

1.2K
Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions
06:18

Author Spotlight: Segmentation and VR for Advanced Neurovascular Interventions

Published on: April 5, 2024

1.1K

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Computer Vision

Background:

  • Deep learning shows promise in left atrial segmentation, with semi-supervised methods often relying on consistency regularization for 3D model training.
  • Existing semi-supervised approaches primarily focus on inter-model consistency, neglecting valuable inter-model discrepancy information.

Purpose of the Study:

  • To develop an improved double-teacher framework that incorporates discrepancy information for enhanced left atrial segmentation.
  • To reduce the computational and data requirements associated with traditional 3D semi-supervised segmentation models.

Main Methods:

  • Designed a double-teacher framework where one teacher model processes 2D information and the other processes both 2D and 3D information.
  • Employed a student model guided by both teacher models.
  • Extracted isomorphic/heterogeneous discrepancy information between student and teacher predictions for optimization, utilizing 3D information to assist 2D models without a full 3D model.

Main Results:

  • The proposed framework achieved excellent performance on the left atrium (LA) dataset.
  • Performance was comparable to leading 3D semi-supervised methods.
  • Addressed limitations of 3D models, such as high memory consumption and limited training data availability.

Conclusions:

  • The novel double-teacher framework effectively utilizes discrepancy information for superior left atrial segmentation.
  • This approach offers a computationally efficient alternative to fully 3D semi-supervised methods, demonstrating strong potential for clinical applications.