Glaucoma Detection Using Enhanced K-Strange Points Clustering Algorithm and Classification

Glaucoma is an eye disorder that majorly affects the optic nerve head in the retina. The damage caused to optic disc leads to gradual loss of peripheral vision which may further result in complete blindness. Glaucoma cannot be cured, hence early and accurate detection is necessary. This paper proposes a method to detect Glaucoma using fundus images. Enhanced K-Strange Points Clustering (EKSTRAP) algorithm is applied to obtain cup, disc and the blood vessels from the Neuro-Retinal Rim (NRR). Further elliptical fitting method is used to compute cup to disc (CDR) ratio. The Inferior-Superior-Nasal-Temporal (ISNT) ratio is obtained using masking. CDR and ISNT are used as inputs to the Naïve Bayes classifier.