Related Experiment Video
Updated: May 4, 2026
09:17
Using Retinal Imaging to Study Dementia
Published on: November 6, 2017
21.7K
A foundation model for generalizable disease detection from retinal images.
Yukun Zhou1,2,3, Mark A Chia4,5, Siegfried K Wagner4,5
1Centre for Medical Image Computing, University College London, London, UK. yukun.zhou.19@ucl.ac.uk.
Nature
|September 13, 2023
Summary
RETFound, a novel foundation model, uses self-supervised learning on unlabelled retinal images to improve AI diagnosis for eye and systemic diseases. This approach enhances model generalizability and reduces expert annotation needs.
Area of Science:
- Ophthalmology
- Medical Artificial Intelligence
- Computer Vision
Background:
- Medical AI shows promise for diagnosing eye and systemic disorders from retinal images.
- Current AI models require extensive annotation and lack generalizability across clinical tasks.
Purpose of the Study:
- To introduce RETFound, a foundation model for retinal images.
- To enable label-efficient adaptation for diverse clinical applications.
- To improve AI model performance and reduce annotation workload.
Main Methods:
- Trained RETFound on 1.6 million unlabelled retinal images using self-supervised learning.
- Adapted the foundation model to specific disease detection tasks using explicit labels.
- Evaluated performance against comparison models in diagnosis, prognosis, and incident prediction.
Main Results:
- Adapted RETFound consistently outperformed comparison models.
- Achieved superior performance with fewer labelled data for various eye and systemic diseases.
- Demonstrated effectiveness in diagnosing sight-threatening eye diseases and predicting heart conditions.
Conclusions:
- RETFound offers a generalizable AI solution for retinal image analysis.
- The model significantly improves diagnostic accuracy and reduces the need for expert annotations.
- Enables broader clinical applications of AI in ophthalmology and beyond.
Related Concept Videos
01:32The Retina
56.6K
The retina is a layer of nervous tissue at the back of the eye that transduces light into neural signals. This process, called phototransduction, is carried out by rod and cone photoreceptor cells in the back of the retina.
56.6K
01:27Diabetic Retinopathy
55
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...
55