Early Detection of Glaucoma using CNN Architecture

The Retinal nerve and the Optic Nerve Fibre Layer (RNFL) are damaged by an eye disease called as Glaucoma, eventually leading to blindness if not treated in time. When this happens, fluid builds up in the retina’s anterior region, which raises eye pressure and damages the optic nerve. Glaucoma results in irreversible vision loss due to increased intraocular pressure (IOP) and optic nerve damage. Ophthalmologists use expensive and time-consuming medical techniques such as HRT and OCT. As a result, to detect glaucoma efficiently and quickly, an automatic computer-aided system is required. The optic cup and optic disc are the primary characteristics that assist in the diagnosis of glaucoma. As a result, segmentation of the optic disc and optic cup properly is critical in detecting the disorder. In this paper, results are obtained using an adaptive method that takes as input raw images, masks, and text files (CDR values). The experimental results are found to be 98% accurate. Furthermore, the proposed method is faster and has a less cost.