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 Videos

A clinically interpretable deep learning pipeline for diabetic retinopathy classification using EfficientNet,

Khirod Kumar Ghadai1, Subrat Kumar Nayak2, Biswa Ranjan Senapati1

  • 1Department of Computer Science and Engineering, Siksha 'O' Anusandhan (Deemed to be) University, Bhubaneswar, Odisha, India.

BMC Ophthalmology
|May 19, 2026
PubMed
Summary

Related Concept Videos

Diabetic Retinopathy01:27

Diabetic Retinopathy

DefinitionDiabetic retinopathy is a microvascular complication of diabetes affecting the retinal blood vessels.Risk FactorsDiabetic retinopathy is present in almost all individuals with type 1 diabetes and more than 60% of those with type 2 diabetes after two decades of disease.The risk increases with poor glycemic control, hypertension, dyslipidemia, smoking, pregnancy, and puberty.Although cataracts and glaucoma are also more frequent in people with diabetes, retinopathy remains the leading...

You might also read

Related Articles

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

Sort by
Same author

AI-assisted Emergency Healthcare using Vehicular Network and Support Vector Machine.

Wireless personal communicationsยท2023
See all related articles

This study developed an interpretable deep learning pipeline for diabetic retinopathy grading, achieving high accuracy. The model shows promise for clinical decision support but requires external validation for advanced disease stages.

Area of Science:

  • Ophthalmology
  • Computer Science
  • Artificial Intelligence

Background:

  • Diabetic retinopathy grading is crucial for preventing vision loss.
  • Deep learning shows promise but faces challenges like class imbalance and interpretability.
  • A clinically interpretable deep learning pipeline for five-class diabetic retinopathy grading was developed.

Purpose of the Study:

  • To develop and evaluate a deep learning pipeline for accurate and interpretable diabetic retinopathy grading.
  • To address challenges in grading, including class imbalance and subtle disease stage differences.
  • To provide a research-stage decision-support tool for ophthalmologists.

Main Methods:

  • An EfficientNet-B1 framework was used with image enhancement (CLAHE), data augmentation (sample-mixing), focal loss, and test-time augmentation.
Keywords:
Data augmentationDeep learningDiabetic retinopathyEfficientNet-B1GradCAM

Related Experiment Videos

  • Model interpretability was assessed using calibration analysis, Grad-CAM, and t-SNE.
  • Performance was evaluated on the APTOS 2019 dataset using accuracy, F1-score, and AUC, with robustness checks via cross-validation.

    • Main Results:

    • The pipeline achieved 84.97% accuracy on the test set with test-time augmentation.
    • The average AUC reached 90.40%, demonstrating strong discriminative ability.
    • The model showed robustness to class imbalance and provided meaningful visual explanations, though severe cases remained challenging.

    Conclusions:

    • The developed pipeline offers competitive performance for diabetic retinopathy grading, combining accuracy with interpretability.
    • The approach is suitable as a research-stage decision-support system.
    • Further external validation is necessary before clinical deployment due to remaining challenges in advanced grading and dataset limitations.