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: Oct 20, 2025

Using Retinal Imaging to Study Dementia
09:17

Using Retinal Imaging to Study Dementia

Published on: November 6, 2017

21.8K

MsTGANet: Automatic Drusen Segmentation From Retinal OCT Images.

Meng Wang, Weifang Zhu, Fei Shi

    IEEE Transactions on Medical Imaging
    |September 14, 2021
    PubMed
    Summary

    Accurate drusen segmentation in retinal OCT images is vital for diagnosing AMD. A new multi-scale transformer network (MsTGANet) and its semi-supervised version (Semi-MsTGANet) improve segmentation accuracy, outperforming existing methods.

    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

    PolyHIPE-derived amine-porphyrin monoliths for mild cycloaddition of low-concentration CO<sub>2</sub>.

    Chemical communications (Cambridge, England)·2026
    Same author

    Bioinformatic analysis of the glutathione S-transferase gene family in Sarcoptes scabiei and protective immunity of Delta-class SsGSTd3 and SsGSTd5 in rabbits.

    Veterinary parasitology·2026
    Same author

    Melatonin Counteracts Mechanical Unloading-Induced Bone Loss Through YTHDF3-Mediated m<sup>6</sup>A Modification of Dapk2 mRNA.

    Journal of pineal research·2026
    Same author

    WRCANet: wavelet residual cross-attention network for meibomian gland segmentation.

    Biomedical optics express·2026
    Same author

    Image Fusion Based on Prior Information.

    IEEE transactions on pattern analysis and machine intelligence·2026
    Same author

    Separation and Characterization of Self-Assembled Nanoparticles from <i>Rheum palmatum</i> L.-<i>Salvia miltiorrhiza</i> Bunge Extract and Their Renoprotective Effects in Acute Kidney Injury.

    Antioxidants (Basel, Switzerland)·2026

    Area of Science:

    • Ophthalmology
    • Medical Imaging
    • Artificial Intelligence

    Background:

    • Drusen are key indicators for diagnosing age-related macular degeneration (AMD) and represent a significant risk factor for its development.
    • Accurate segmentation of drusen in optical coherence tomography (OCT) images is critical for the early detection of AMD.
    • Challenges in drusen segmentation include variations in size and shape, blurred boundaries, speckle noise, and a scarcity of pixel-level annotated OCT datasets.

    Purpose of the Study:

    • To propose a novel multi-scale transformer global attention network (MsTGANet) for accurate drusen segmentation in retinal OCT images.
    • To introduce a semi-supervised version (Semi-MsTGANet) utilizing pseudo-labeled data augmentation to address the limited availability of annotated data.
    • To enhance the model's capability in capturing multi-scale features and global contextual information for improved 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

    3.0K
    Application of Optical Coherence Tomography to a Mouse Model of Retinopathy
    08:22

    Application of Optical Coherence Tomography to a Mouse Model of Retinopathy

    Published on: January 12, 2022

    4.6K

    Related Experiment Videos

    Last Updated: Oct 20, 2025

    Using Retinal Imaging to Study Dementia
    09:17

    Using Retinal Imaging to Study Dementia

    Published on: November 6, 2017

    21.8K
    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

    3.0K
    Application of Optical Coherence Tomography to a Mouse Model of Retinopathy
    08:22

    Application of Optical Coherence Tomography to a Mouse Model of Retinopathy

    Published on: January 12, 2022

    4.6K

    Main Methods:

    • Development of MsTGANet, a U-shaped architecture incorporating a multi-scale transformer non-local (MsTNL) module for capturing long-range dependencies.
    • Integration of a multi-semantic global channel and spatial joint attention module (MsGCS) to facilitate feature fusion and contextual learning.
    • Implementation of a semi-supervised approach (Semi-MsTGANet) leveraging unlabeled data through pseudo-labeling to boost performance.

    Main Results:

    • MsTGANet effectively captures multi-scale non-local features and global contextual information.
    • Semi-MsTGANet demonstrates improved segmentation performance by effectively utilizing unlabeled data.
    • Both proposed methods achieved superior segmentation accuracy compared to state-of-the-art Convolutional Neural Network (CNN)-based methods in comprehensive experiments.

    Conclusions:

    • The proposed MsTGANet and Semi-MsTGANet offer significant advancements in automated drusen segmentation from retinal OCT images.
    • These novel deep learning approaches address key challenges, including data scarcity and complex image features.
    • The findings suggest a promising direction for improving early AMD diagnosis through enhanced OCT image analysis.