Abstract: Glaucoma is a disease in which the optic nerve of the eye gets destroyed. As a result, it causes vision loss or blindness. However, with earlier diagnosis and treatment, eyes can be protected against severe vision loss. Most of times peripheral vision can be damaged earlier than an individual’s central vision by Glaucoma because it does not show any sign and symptoms. The existing procedures to detect Glaucoma are time consuming and uncertain at the clinic. We propose a low cost Glaucoma detection system which is a computer-based technology and therefore, it uses algorithms to instantaneously detect and classify healthy and Glaucoma eye. It does this by analysing Region of Interest (ROI) of images through implementation of various image extraction features like Colour histogram, Haralick texture features using GLCM matrix, Multi level wavelet based feature using discrete wavelet transform. For Classification of healthy and Glaucoma eye we proposed Supervised Machine Learning approach using Random forest algorithm. The performance of the proposed method was evaluated in terms of accuracy, specificity, sensitivity and more parameters. From the experimental results of the proposed system, the accuracy, specificity, sensitivity is obtained as 95.65, 96.66 and 95.16 respectively.

Keywords: Glaucoma, GLCM matrix, Wavelet, Random Forest.

| DOI: 10.17148/IJIREEICE.2020.8626