Abstract
The rapid evolution of ophthalmic imaging—ranging from optical coherence tomography (OCT) and confocal microscopy to wavefront aberrometry and ocular surface interferometry—has generated unprecedented volumes of high-dimensional optical data. However, the diagnostic potential of these modalities remains underutilized due to interpretation bottlenecks, variability in clinical expertise, and limited transparency of conventional deep learning systems. This study introduces an Explainable Artificial Intelligence (XAI) framework designed to enhance the reliability, interpretability, and clinical applicability of AI-based ophthalmic diagnostics. By integrating optical physics–informed feature extraction, attention-based visualization, and knowledge-guided interpretation modules, the proposed system enables transparent analysis of retinal microstructures, corneal topography, tear film dynamics, and laser-induced tissue responses. We demonstrate the framework’s ability to identify early pathological biomarkers, support refractive surgery planning, and provide interpretable recommendations for ocular surface and retinal disease management. Visual reasoning maps, causal evidence traces, and human-aligned explanations promote clinician trust and facilitate safe AI adoption in precision eye care. This work highlights the transformative potential of XAI in bridging optical imaging innovation with next-generation ophthalmic diagnostics, contributing to the broader vision of intelligent, transparent, and patient-centered eye health technologies.
Biography
Mini Han Wang completed her PhD in Ophthalmology and Visual Sciences from the Chinese University of Hong Kong and a second PhD in Data Science from the City University of Macau. She is an Sinore Research Fellow and Senior AI Engineer at the Chinese Academy of Sciences–affiliated Zhuhai Institute of Advanced Technology Chinese Academy of Sciences. She currently serves as President of the Macau Digital Healthcare and Artificial Intelligence Association. Dr. Wang has published over 60 peer-reviewed papers and holds more than 50 invention patents. She also serves as Guest Editor or Editorial Board Member for several international journals in AI, ophthalmology, and digital health.
