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: May 21, 2026

Simultaneous Eye Tracking and Single-Neuron Recordings in Human Epilepsy Patients
07:43

Simultaneous Eye Tracking and Single-Neuron Recordings in Human Epilepsy Patients

Published on: June 17, 2019

A gated task-attentive multi-task network for unified retinal image analysis.

Muhammad Zaheer Sajid1, Imran Qureshi2, Muhammad Fareed Hamid3

  • 1Department of Electrical and Computer Engineering, George Mason University, Fairfax, USA. msajid4@gmu.edu.

Scientific Reports
|May 19, 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

The Cutting Edge: Interventional Strategies for Pain Management.

Physical medicine and rehabilitation clinics of North America·2026
Same author

TriDermCancerNet: A hybrid deep learning framework for skin cancer classification.

The Journal of international medical research·2026
Same author

Decoding basal ganglia motor circuit dysfunction from handwriting: a physics-informed neural signal interpretation framework for Parkinson's disease screening.

Frontiers in neuroinformatics·2026
Same author

Lightweight and Energy-Aware Intrusion Detection for Industrial IoT Using TinyML and Edge AI.

Scientific reports·2026
Same author

Privacy-aware distributed intelligence with tokenized trust for low-latency task offloading in 6G vehicular edge networks.

Scientific reports·2026
Same author

Deep learning for automatic segmentation of the inferior alveolar nerve using a hybrid CNN-transformer framework.

Scientific reports·2026
This summary is machine-generated.

A new deep learning model, GTAM-Net, jointly analyzes retinal images for optic disc segmentation and diabetic retinopathy (DR) grading. This integrated approach improves accuracy and stability for large-scale screening, combating preventable blindness.

Area of Science:

  • Ophthalmology
  • Medical Imaging
  • Artificial Intelligence

Background:

  • Diabetic retinopathy (DR) is a leading cause of preventable blindness globally.
  • Current deep learning methods often analyze optic disc segmentation and DR grading separately, hindering the use of shared contextual information.
  • Accurate and efficient large-scale screening tools for DR are urgently needed.

Purpose of the Study:

  • To propose GTAM-Net, a novel Gated Task-Attentive Multi-Task Network for simultaneous optic disc segmentation and DR severity grading.
  • To leverage shared anatomical and contextual cues between tasks within a single end-to-end network.
  • To improve the accuracy and stability of automated retinal image analysis for DR screening.

Main Methods:

  • GTAM-Net employs a gated task-attentive block to dynamically share features between segmentation and grading tasks.
Keywords:
Diabetic retinopathy gradingGated attention networkMulti-task learningOptic disc segmentationRetinal image analysis

More Related Videos

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss
07:12

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss

Published on: April 11, 2025

Related Experiment Videos

Last Updated: May 21, 2026

Simultaneous Eye Tracking and Single-Neuron Recordings in Human Epilepsy Patients
07:43

Simultaneous Eye Tracking and Single-Neuron Recordings in Human Epilepsy Patients

Published on: June 17, 2019

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss
07:12

A Gaze-Contingent Display Framework for Perceptual Learning Research with Simulated Central Vision Loss

Published on: April 11, 2025

  • A multi-scale feature pyramid is utilized to preserve hierarchical contextual information.
  • Uncertainty-based loss weighting is implemented to prevent task domination during training.

    • Main Results:

    • The model achieved up to 98.17% Dice score for optic disc segmentation and 99.12% accuracy for DR grading.
    • Performance was competitive across five diverse public datasets (IDRiD, DDR, Messidor-2, APTOS, REFUGE).
    • Cross-dataset evaluations demonstrated notable stability under varying imaging conditions.

    Conclusions:

    • The proposed multi-task learning approach (GTAM-Net) is effective and stable for joint retinal image analysis.
    • This method offers a promising solution for large-scale diabetic retinopathy screening pipelines.
    • Integrating segmentation and grading tasks enhances performance by utilizing complementary information and avoiding negative transfer.