Retinal Disease Detection using CNN through Optical Coherence Tomography Images

Abstract

The thin-layered tissue retina is an important part of the eye. Diabetic Macular Edema, Drusen, and Choroidal Neovascularization are major maladies of the retina. Optical coherence tomography (OCT) is used to take cross-sectional pictures of the retina and diagnose these maladies. Retinal disease diagnosis needs an extent accuracy for working as petty errors in diagnosis can result in catastrophic repercussions. The ultimate goal of this paper is to detect retinal disease at right time and help to combat the increasing number of cases by prompt treatment. The objective of the paper is to find the most optimal CNN model for diagnosing retinal disease, by implementing CNNs with different optimizers and regularization conditions on a categorical dataset. The selected architecture can be a useful tool for physicians and for the medical community for correctly identifying retinal infection with accurate and prompt treatment.