Diabetic Retinopathy (DR) is a major problem that impairs anthropological eyesight. It is usually formed at the rear side of the eye retina, when light-sensitive tissues' blood vessels are damaged. Based on symptoms or diseases in the eye, such as mild visual difficulties, vision loss, blindness, damage to blood vessels in the retina caused by high blood sugar levels, etc., attention is required. Investigators have confronted a crucial test for the early identification of eye illness using image recognition methods for many years. A method for automatically identifying diabetic retinopathy based on topographies such as point-based from retinal pictures utilizing SVM and basket of topographies techniques. There are four steps in this procedure: The input eye retina picture is pre-processed in the first step. We smear re-scaling geometric changes like revolution for picture enhancement in the second phase. In the third stage, we present an automated technique based on point-based features and a list of pictures for first diabetes or non-diabetic diagnosis utilizing SVM and a basket of topographies such as Haar modification image processing methods. Lastly, we liken the measurements like RMSE, PSNR, Elapsed Time and Accuracy utilizing this future method.