Deep Learning-Based Classification of Retinal Pathologies

Age-related Macular Degeneration (AMD), Diabetic Macular Edema (DME), Epiretinal Membrane (ERM), Retinal Artery Occlusion (RAO), Retinal Vein Occlusion (RVO), and Vitreomacular Interface Disease (VID) are prevalent eye conditions that can lead to partial or complete vision impairment and blindness. Addressing these challenges in eye care necessitates advanced imaging technologies like Optical Coherence Tomography (OCT). The evolution of OCT from time-domain to frequency-domain techniques has significantly enhanced its utility in routine clinical procedures. This paper introduces a novel R50-CapsNet architecture designed to classify retinal diseases more accurately and reliably. Our approach aims to improve diagnostic accuracy for the OCTDL and Kermany datasets.