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

Vision-language models for human motion understanding: Lessons from stroke rehabilitation.

PLOS digital health·2026
Same author

Multi-modal AI for comprehensive breast cancer prognostication.

Nature communications·2026
Same author

Leveraging Fine-Tuned Large Language Models for Interpretable Pancreatic Cystic Lesion Feature Extraction and Risk Categorization.

AJR. American journal of roentgenology·2026
Same author

Revealing information from weak signals in electron energy-loss spectroscopy with a a deep denoiser.

Micron (Oxford, England : 1993)·2025
Same author

An interpretable and adaptive autoencoder for efficient tissue deconvolution.

Nucleic acids research·2025
Same author

Visualizing nanoparticle surface dynamics and instabilities enabled by deep denoising.

Science (New York, N.Y.)·2025

Related Experiment Video

Updated: Oct 24, 2025

Author Spotlight: Unraveling the Pathogenesis of Age-Related Macular Degeneration and Discovering Potential Therapies
06:16

Author Spotlight: Unraveling the Pathogenesis of Age-Related Macular Degeneration and Discovering Potential Therapies

Published on: July 28, 2023

2.8K

ANALYSIS OF TRANSFER LEARNING FOR SELECT RETINAL DISEASE CLASSIFICATION.

Rony Gelman1, Carlos Fernandez-Granda1,2

  • 1Courant Institute of Mathematical Sciences, New York University, New York, New York; and.

Retina (Philadelphia, Pa.)
|August 16, 2021
PubMed
Summary

Transfer learning significantly improved diabetic retinopathy (DR) classification using fundus photography. However, its effect on spectral domain optical coherence tomography (SD-OCT) classification of retinal diseases was minimal.

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
Development of a Gaze-Contingent Display Framework Designed for Perceptual and Oculomotor Research with Simulated Central Vision Loss
07:12

Development of a Gaze-Contingent Display Framework Designed for Perceptual and Oculomotor Research with Simulated Central Vision Loss

Published on: April 11, 2025

610

Related Experiment Videos

Last Updated: Oct 24, 2025

Author Spotlight: Unraveling the Pathogenesis of Age-Related Macular Degeneration and Discovering Potential Therapies
06:16

Author Spotlight: Unraveling the Pathogenesis of Age-Related Macular Degeneration and Discovering Potential Therapies

Published on: July 28, 2023

2.8K
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
Development of a Gaze-Contingent Display Framework Designed for Perceptual and Oculomotor Research with Simulated Central Vision Loss
07:12

Development of a Gaze-Contingent Display Framework Designed for Perceptual and Oculomotor Research with Simulated Central Vision Loss

Published on: April 11, 2025

610

Area of Science:

  • Ophthalmology
  • Medical Imaging
  • Artificial Intelligence

Background:

  • Diabetic retinopathy (DR) and other retinal diseases pose significant challenges in diagnosis and management.
  • Accurate classification of these conditions is crucial for timely intervention and patient outcomes.
  • Deep learning models show promise in analyzing medical images for disease detection.

Purpose of the Study:

  • To evaluate the impact of transfer learning on the performance of deep neural networks for classifying diabetic retinopathy (DR) from fundus photography.
  • To assess the effect of transfer learning on classifying retinal diseases such as drusen, choroidal neovascularization, and diabetic macular edema using spectral domain optical coherence tomography (SD-OCT) images.

Main Methods:

  • Five open-source deep neural networks and four custom CBR networks were trained and evaluated.
  • Two tasks were performed: DR classification using fundus images and retinal disease classification using SD-OCT images.
  • Performance was measured by Kappa coefficient for DR and accuracy for SD-OCT, comparing models with and without transfer learning.

Main Results:

  • Transfer learning consistently improved the Kappa coefficient for DR classification across all tested networks, with increases ranging from 0.152 to 0.556.
  • For SD-OCT classification, transfer learning led to accuracy increases in four of five open-source networks (1.8%-3.5%) but showed minimal or negative effects on the remaining networks.
  • Eight of nine networks demonstrated minimal impact of transfer learning on SD-OCT-based classification accuracy.

Conclusions:

  • Transfer learning substantially enhances the performance of deep learning models for diabetic retinopathy classification from fundus photography.
  • The application of transfer learning yields minimal performance gains for classifying retinal diseases using SD-OCT imaging across most evaluated networks.
  • These findings suggest a differential benefit of transfer learning depending on the imaging modality and specific classification task in ophthalmology.