Özet:

Abstract In recent years, diabetes rates are increasing drastically due to elevated blood sugar in the human body. With the increase in diabetes rate, it impacts the eye it requires regular examination to prevent blindness. Eye diseases affects the person with a higher glucose rate in the blood. However, after certain duration blood sugar remains in the retina and affects the retina lead to damage in the eye. The presence of blood glucose in the vessels of the eye damages the eye vessels and causes leakage of fluid. Eye diseases’ impact on working-age adults causes the loss of eyesight. Even though treatment can help but early intervention prevents loss of vision due to eye diseases such as Diabetic Retinopathy (DR), maculopathy, Glaucoma, Exudates, and Hemorrhage. This paper proposed an Entropy segmentation Survival Analysis Optimization (EsSO) for the classification of Diabetic Retinopathy (DR), maculopathy, Glaucoma, Exudates, and Hemorrhage. The proposed architecture performs segmentation based on the estimation of entropy. The feature extraction and classification are performed with the optimization of the GLCM features in the images. To perform image optimization GLCM features with the black widow are implemented. Through computed feature classification is performed with the conventional neural network model for classification. The classification is performed for estimation of different diseases in the eye Diabetic Retinopathy (DR), maculopathy, Glaucoma, Exudates and Hemorrhage. The proposed EsSO model concentrated highly on the intervention of eye diseases for diagnosis and treatment. The performance of the developed model is comparatively examined with the conventional technique. The proposed EsSO model provides an accuracy of 96% whereas the conventional classifiers SVM and RF provides the accuracy of 91% and 94% respectively. The evaluation expressed that the proposed EsSO model exhibits ~4% improvement than the conventional classifiers.

Anahtar Kelimeler: