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 Experiment Video

Updated: Jun 27, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Semi-SwinUNeTR: Towards 3D Swin Vision Transformer-Based UNet for Medical Image Segmentation with Limited

Yinbing Tian1,2,3,4, Ziyang Wang5, Li Guo1,2,3,4

  • 1School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing 100876, China.

Bioengineering (Basel, Switzerland)
|June 26, 2026
PubMed
Summary

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

TriConvUNeXt: A Pure CNN-Based Lightweight Symmetrical Network for Biomedical Image Segmentation.

Journal of imaging informatics in medicine·2024
Same author

Unveiling breast cancer risk profiles: a survival clustering analysis empowered by an online web application.

Future oncology (London, England)·2023
See all related articles

This study introduces Semi-SwinUNeTR, a novel semi-supervised framework for 3D brain tumor segmentation using limited MRI data. The method enhances accuracy by combining model and data consistency, improving diagnosis and treatment planning.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Computational Biology

Background:

  • Accurate brain tumor segmentation in MRI is crucial for clinical applications but challenged by irregular tumor shapes and limited annotated data.
  • Deep learning models struggle with heterogeneous tumor patterns and ambiguous boundaries, necessitating efficient methods for limited-label scenarios.

Purpose of the Study:

  • To develop a semi-supervised framework, Semi-SwinUNeTR, for 3D brain tumor segmentation using limited annotated magnetic resonance imaging (MRI) data.
  • To enhance segmentation performance by integrating a SwinUNeTR backbone with a dual-consistency learning strategy and voxel-wise consistency weighting.

Main Methods:

  • Utilized SwinUNeTR, a transformer-based network with shifted-window self-attention, for hierarchical volumetric representation learning in a 3D segmentation task.
Keywords:
SwinUNeTRauxiliary weightingbiomedical image analysisinterpolation consistencysemi-supervised learningvision transformervolumetric medical image segmentation

More Related Videos

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

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

Related Experiment Videos

Last Updated: Jun 27, 2026

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

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

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
04:23

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

  • Implemented a dual-consistency semi-supervised learning strategy, incorporating mean teacher model consistency and interpolation-based data consistency.
  • Introduced voxel-wise consistency weights to prioritize supervision in complex tumor regions and along irregular boundaries.

    • Main Results:

    • Semi-SwinUNeTR achieved strong performance across various labeled data ratios on the BraTS 2019 benchmark, with Dice scores ranging from 84.93% to 87.83%.
    • The weighted consistency extension further improved Dice scores to 88.59% and reduced HD95 to 7.4533 at 80% labeled data.
    • Demonstrated significant improvements in 3D brain tumor segmentation accuracy with limited annotations.

    Conclusions:

    • The proposed Semi-SwinUNeTR framework effectively addresses the challenge of limited annotations in 3D brain tumor segmentation.
    • Combining SwinUNeTR with dual-consistency learning and voxel-wise weighting offers a robust strategy for semi-supervised medical image segmentation.
    • The findings highlight the potential of advanced deep learning techniques for improving computer-assisted diagnosis and treatment planning in neuro-oncology.