Chief Supervisor

Prof. HE Mingguang

Project Title

Synopsis

Ophthalmic diseases exert a significant impact on global health and quality of life, relying heavily on clinical records and various forms of medical imaging for diagnosis and treatment decisions. The highly specialized and multimodal nature of ophthalmology poses challenges due to enormous demand, financial constraints, and a shortage of professionals. Recent advancements in artificial intelligence (AI), particularly large language models (LLMs), have introduced powerful tools for healthcare professionals, offering rich external knowledge and robust reasoning capabilities.
This project aims to leverage the comprehensive multimodal ophthalmic data and advanced algorithms to develop and validate a unified ophthalmic AI system empowered by LLMs. This system will integrate and combine diverse sources of multimodal ophthalmic data, handling multiple downstream ophthalmic tasks including disease classification, diagnosis, medical report generation and visual question answering.
Prospective studies involving real-world clinical doctors and patients will be conducted to further evaluate the clinical effectiveness and safety of the system. The successful development and validation of this system will significantly enhance diagnostic accuracy for physicians, improve patients' access to care, and revolutionize diagnosis, treatment, and disease management in ophthalmology.