Chief Supervisor

Prof. HE Mingguang

Project Title

Synopsis

Eye digital twins (EDTs) are computational tools designed to capture key functional and anatomical characteristics of a patient’s eyes. They enable detailed investigation of diseases and help predict responses to therapy. Moving beyond conventional imaging, EDTs integrate and synchronize multi-modal data, serving as a unified foundational data engine. This research explores how such rich data streams can facilitate autonomous knowledge discovery through the integration of advanced deep learning and generative AI techniques. Deep neural networks will be applied to perform complex feature extraction, pattern recognition, and anomaly detection within the digital twin’s dataset. Additionally, generative AI models will be used to simulate unseen scenarios, predict future states, and impute missing information—thereby enhancing the digital twin’s fidelity and revealing latent insights. The outcome aims to deliver a transformative framework that elevates visual data analysis from passive observation to active, predictive knowledge generation. This holds significant potential for applications in autonomous systems, predictive maintenance, and scientific discovery.