A novel hybrid approach to blaze out a new path for glaucoma detection, monitoring and sustainable results in fundus images

摘要

Glaucoma can be metaphorically called gruesome thief who steals away the most precious eyesight slowly, silently and forever which often remains a frightening prospect to patients and frustrates ophthalmologist. It is a persisting disease and the advancement of which can be routed out if diagnosed in the beginning. Here a novel hybrid method for automated detection of glaucoma is being proposed, which has the ability to support the ophthalmologists in disease detection in the early stage. The algorithm proposed here takes a pre-processed fundus image and calculates CDR from extracted optic cup and optic disc. The suggested system applies hybrid features for conformation of training and testing the glaucoma image, SVM classifier for detection of glaucoma. The proposed system is estimated by utilising DrishtiGSl database. The sensitivity, specificity and accuracy obtained in our methods are being compared with other methods in the literature survey. The proposed system guarantees to identify glaucoma cases with 100% precision there by making it befitting to claim as an automated medical tool for diagnosing glaucoma.