Machine learning based approach for lesion segmentation and severity level classification of diabetic retinopathy

Abstract

Diabetic Retinopathy (DR) is considered as the most significant factor that results in retinal damage which leads to eye issues and further serious vision loss. Person suffering from high blood glucose level or having diabetes are prone to diabetic retinopathy. At initial stage prominent symptoms are not seen. Day by day it progresses to serious level and leads to only option that is operation. Operation is scary phenomenon in terms of money and consumes ample time.  In the present work a system is designed for DR diagnosis and its classification. The system attains timely and accurate detection of DR and its concerned symptoms. Appropriate lesions are selected and prominent features are extracted from it which results in better DR classification. Student Feedback Artificial Tree Optimization (SFATO) methodology is used to carry out the present research work. The SFATO attains 91% Accuracy, 92% of Sensitivity and 90.5% of Specificity which is better than existing systems.

URN:NBN:sciencein.jist.2023.v11.576