Glaucoma Detection and Classification

Glaucoma is a chronic eye illness that results from optic nerve damage caused by high intraocular pressure. There are no symptoms of glaucoma in the early stages, but as the disease continues, it can lead to irreversible blindness. Diagnosis of glaucoma in the clinical environment includes intraocular pressure measurement, visual field testing, or examination of the optical disk of fundus images. In large-scale screening scenarios, these manual assessments are not precise, mostly in developing countries due to the insufficiency of trained experts and scarce modern imaging equipment. In this paper, several models are being used to study glaucoma detection. The models chosen are: VGG19, VGG19+LSTM, InceptionV3, and InceptionV3+LSTM. Every model is being worked with K-fold cross-validation and data augmentation to overcome the limitation of a small dataset. The features extracted are used to classify the input image and are then projected to be either glaucomatous or normal. Finally, the values obtained for various performance evaluation parameters are compared.