DIABETIC RETINOPATHY DETECTION USING TRANSFER LEARNING BASED DEEP LEARNING MODEL
Keywords:
Convolutional Neural Network(CNN), DeepLearning(DL), Optimizer, Transfer learning, Augmentation, Adaptive Moment Estimation(ADAM), Root Mean Square Propagation(RMSprop), Stochastic Gradient Descent with Momentum(SGDM), MoblieNet-V2, GoogleNet, ResNet 18, Diabetic Retinopathy (DR)Abstract
Diabetic Retinopathy (DR), an eye condition that mostly affects people with diabetes, is one of the main causes of adult blindness. Eighty percent of patients with DR have been shown to have chronic diabetes over a period of 10 to 15 years. As the disease progresses, it may result in a permanent loss of vision. Early detection allows for an early diagnosis of DR. This work focuses mostly on applying deep learning (DL) and transfer learning approaches to diagnose diabetic retinopathy. The Messidor 2 dataset, which includes 1748 macula-centered eye fund images, of which 1744 are labeled and 4 are not labeled, is considered in this study. There are two classes in the labeled data: DR0 (fundus image without diabetic retinopathy) and DR1 (fundus image with diabetic retinopathy). Using the principles of transfer learning, the pre-trained models ResNet18, MobileNet-V2, and GoogleNet are refined and optimized via three distinct optimizers: ADAM, RMSprop, and SGDM. The developed models provide an accuracy of 98.9% for ResNet18 models, 74.4% for MobileNet-V2, and 68.1% for GoogleNet. The developed models show better performance than the existing algorithms.
