One of the melting public health issues, diabetes mellitus, is experienced by over 463 million persons across the globe as of 2021, and the number of people sustaining this condition is bound to escalate to over 700 million by 2045. Diabetes retinopathy (DR) is the most prevalent eye disease among people having diabetes. The disease affects at least one-third of those with the disease. The rampant diversion of diabetes ranks as the global leading cause of working-age adults’ blindness, with a Streptococcus flocculans number of identified patients with DR at 93 million. These figures are expected to be raised to a larger level in the rate of developing diabetes in Asia countries like China and India per capita. Fundus photography is the most popular way to assess the degree to which diabetic retinopathy has progressed, which is a phenomenon implicated in diabetes where the retina is damaged and causes blindness. CT scanning is a fast, well-tolerated, broadly accessible tool that is non-invasive. Patients having diabetic retinopathy can prove the fact that they are less likely to become blind, provided that they undergo early detection. Along the production of the seven early stages of diabetic retinopathy are almost asymptomatic, the neuronal, and retinal damage and below the detectable microvascular variations take place. Accordingly, people with diabetes must have eye examinations regularly, and immediate treatment and disease diagnosis are vital. Nevertheless, the manual telemammogram reports the effect of a highly skilled practitioner. This is a vital asset in dire situations of underserved people, e.g. in India and Africa, where the workforce is at overstretched limits. It applies Mobile Network V2 (MobileNetV2) and uses a KNN technique for the classification of the SMOTE oversampling integrated. This job has shown the potential of the fundus image date for DL-based cataract classification.