VisionUnite: A Vision-Language Foundation Model for Ophthalmology Enhanced with Clinical Knowledge

TL;DR

VisionUnite is a comprehensive vision-language model for ophthalmology that integrates clinical knowledge, trained on large datasets, and demonstrates diagnostic and educational capabilities comparable to junior ophthalmologists.

Contribution

Introduction of VisionUnite, a novel ophthalmology-focused vision-language foundation model enhanced with clinical knowledge and trained on extensive datasets, outperforming existing models in diagnostics and education.

Findings

Outperforms GPT-4V and Gemini Pro in diagnostics.

Demonstrates diagnostic capabilities similar to junior ophthalmologists.

Effective in clinical scenarios including multi-disease diagnosis and patient interaction.

Abstract

The need for improved diagnostic methods in ophthalmology is acute, especially in the underdeveloped regions with limited access to specialists and advanced equipment. Therefore, we introduce VisionUnite, a novel vision-language foundation model for ophthalmology enhanced with clinical knowledge. VisionUnite has been pretrained on an extensive dataset comprising 1.24 million image-text pairs, and further refined using our proposed MMFundus dataset, which includes 296,379 high-quality fundus image-text pairs and 889,137 simulated doctor-patient dialogue instances. Our experiments indicate that VisionUnite outperforms existing generative foundation models such as GPT-4V and Gemini Pro. It also demonstrates diagnostic capabilities comparable to junior ophthalmologists. VisionUnite performs well in various clinical scenarios including open-ended multi-disease diagnosis, clinical explanation, and patient interaction, making it a highly versatile tool for initial ophthalmic disease screening. VisionUnite can also serve as an educational aid for junior ophthalmologists, accelerating their acquisition of knowledge regarding both common and underrepresented ophthalmic conditions. VisionUnite represents a significant advancement in ophthalmology, with broad implications for diagnostics, medical education, and understanding of disease mechanisms. The source code is at https://github.com/HUANGLIZI/VisionUnite.