Dry eye disease is a major ophthalmic health issue that severely affects the visual function and quality of life of the national population. Its core pathological mechanism involves a vicious cycle of ocular surface inflammation and neural abnormalities; however, clinical practice still lacks an objective and quantitative gold standard for diagnosis. Although in vivo confocal microscopy (IVCM) enables cellular-level, in vivo observation of the ocular surface, image analysis remains heavily dependent on manual interpretation, highlighting an urgent need for an intelligent quantitative framework.This project aims to construct a high-quality, standardized ocular surface imaging database and develop a high-precision deep learning algorithm to achieve accurate segmentation and quantification of corneal nerves (including both whorl-like and linear patterns) and inflammatory cells, and to validate their associations with clinical indicators of dry eye disease. The ultimate goal is to develop and evaluate an IVCM-based multimodal intelligent diagnostic system for dry eye, transforming IVCM from an observational tool into an intelligent decision-support system, with real-world performance validated through an independent prospective cohort.This project is expected to establish a multimodal AI diagnostic model for dry eye, create a standardized computational framework for imaging biomarkers, and enable a paradigm shift from qualitative description to quantitative diagnosis. The findings will provide reliable decision-making support for precision subtyping and personalized treatment of dry eye disease, advancing ophthalmic practice toward a data-driven, intelligent paradigm.