Date of Publication :25th July 2024

Abstract:One of the main complications of diabetes and the primary cause of blindness worldwide is diabetic retinopathy or DR. To prevent vision loss, early identification of diabetic retinopathy and related ocular diseases such cataracts, glaucoma, and macular edema is crucial. In this work, we describe an automated approach that uses fundus images to identify and classify DR and related ocular illnesses. We first extract features from fundus images using the VGG16 convolutional neural network (CNN) architecture, then for multi-class categorization, we employ an Extra Trees classifier. We established a remark- able 91.43% accuracy through experimental validation on a customized dataset from Kaggle, proving the efficacy of our method in detecting DR and related ocular disorders. Automated systems like ours hold great potential for enhancing healthcare outcomes by facilitating efficient screening and diagnosis of diabetic retinopathy and associated eye conditions.

Reference :