RET-CLIP: A Retinal Image Foundation Model Pre-trained with Clinical Diagnostic Reports

TL;DR

RET-CLIP is a novel retinal image foundation model trained on a large clinical dataset, demonstrating superior performance across multiple ophthalmic diagnostic tasks and addressing data scarcity in medical vision-language modeling.

Contribution

The paper introduces RET-CLIP, a retinal image foundation model trained with a tripartite strategy on nearly 200,000 patients, enhancing generalization in ophthalmic diagnosis.

Findings

Outperforms existing benchmarks in eight datasets

Effective in diagnosing diabetic retinopathy, glaucoma, and multiple diseases

Demonstrates strong generality across diverse ophthalmic diagnostic tasks

Abstract

The Vision-Language Foundation model is increasingly investigated in the fields of computer vision and natural language processing, yet its exploration in ophthalmology and broader medical applications remains limited. The challenge is the lack of labeled data for the training of foundation model. To handle this issue, a CLIP-style retinal image foundation model is developed in this paper. Our foundation model, RET-CLIP, is specifically trained on a dataset of 193,865 patients to extract general features of color fundus photographs (CFPs), employing a tripartite optimization strategy to focus on left eye, right eye, and patient level to reflect real-world clinical scenarios. Extensive experiments demonstrate that RET-CLIP outperforms existing benchmarks across eight diverse datasets spanning four critical diagnostic categories: diabetic retinopathy, glaucoma, multiple disease diagnosis, and multi-label classification of multiple diseases, which demonstrate the performance and generality of our foundation model. The sourse code and pre-trained model are available at https://github.com/sStonemason/RET-CLIP.