Publication

Abstract : The progressive deterioration of the optic nerve, which results in irreversible vision loss if left untreated, is the hallmark of a group of eye illnesses known as glaucoma. High intraocular pressure (IOP) is a significant risk factor, but normal IOP is not always necessary for the development of normal-tension glaucoma (NTG). Even so, glaucoma can occur in the absence of high blood pressure. Convolutional neural networks (CNNs) [6], in particular, are deep learning models that have recently demonstrated potential in the diagnosis of glaucoma. These models are capable of detecting minute patterns suggestive of glaucoma by examining massive collections of retinal fundus images. The efficacy of the Inception V3 model [20] in automated glaucoma diagnosis was demonstrated by its 88 % accuracy in separating glaucomatous from non-glaucomatous images after it was trained with a large amount of data and enhanced methodologies.