Retinal Glaucoma Detection from Digital Fundus Images using Deep Learning Approach

Abstract

A disorder called glaucoma that damages the optic nerve can result in either a partial or whole loss of vision. As a reason, it is critical to start glaucoma screening at a young age. Glaucoma symptoms do not manifest until the condition is advanced and the patient has already experienced considerable vision loss. The bulk of early diagnostic methods rely on careful feature engineering. Fundus images are particularly useful in the clinical context for the early detection of vision problems.Because of its superior performance, In related fields including image synthesis, disease segmentation, biomarker segmentation, and illness identification, deep learning is being employed more and more often. Convolutional neural networks have lately been used to diagnose glaucoma and other eye problems in ophthalmology. They have been effective in the early diagnosis of several disorders. Many layers of highly connected neural networks. The study approach makes use of models that were previously trained on ImageNet using the dristi dataset with the accuracy of almost 97% that was achieved, it is evident that the built classification model can accurately diagnose glaucoma.