Abnormality Estimation of Conjunctival Hyperemia Using Smartphone Camera and Deep Learning

A method for quantifying the degree of conjunctival hyperemia through smartphone photography for self-medication is introduced. When a ring light is attached to a smartphone and the eye is photographed, it becomes possible to capture clear images of the blood vessels in the eye. We estimate the degree of hyperemia for a new eye image by employing the skip-GANomaly deep learning model on a large dataset of images depicting normal eyes with relatively minimal redness. This system enables convenient monitoring of hyperemia levels via smartphone at any time, significantly enhancing self-medication practices.