A Generalized Deep Learning Model for Glaucoma Detection

Abstract

Glaucoma is an ocular condition that may lead to blindness at its advanced stages. It results due to damaged optic nerve and does not have many early symptoms. Ophthalmologists may use perimetry, tonometry, and ophthalmoscopy to diagnose a glaucomatous eye. Computer-aided detection of glaucoma has also been possible using deep learning. This paper presents a generalized deep learning model for glaucoma detection using fundus images. Unlike previous studies, the model is trained and tested on multiple datasets and architectures. The results show that the model is comparable or better than earlier work in the literature 80% of the time.