GLAUCOMA DETECTION FROM FUNDUS IMAGES

Abstract: The suggested glaucoma diagnostic system incorporates segmentation of the eye nerve disc and cup of the visual nerve utilizing the U-Net framework, alongside glaucoma classification via the VGG16 model. This system aims to improve precision and effectiveness in detecting glaucoma, facilitating prompt treatment for patients. The system will utilize the U-Net framework to delineate the optic cup and disc areas inside retinal pictures. The VGG16 framework will be Utilized for two-class of grouping, glaucoma status, using the segmented optic regions as input. This model will discern between instances where glaucoma and those without. Clinical Use: The system is designed to support healthcare practitioners in precisely identifying glaucoma in its early stages. It serves as a diagnostic tool, offering clinician’s dependable information to enhance their decision-making process. The system requires labeled datasets of retinal images for training and evaluation. These datasets must include annotations for optic disc, optic cup, and glaucoma status.