Detection of Glaucoma with Deep Learning

Abstract

Glaucoma is an ocular disorder caused due to increased uid pressure in the optic nerve. It damages the optic nerve subsequently and causes loss of vision. The available scanning methods are Heidelberg Retinal Tomography (HRT), Scanning Laser Polarimetry (SLP) and Optical Coherence Tomography (OCT). These methods are expensive and require experienced clinicians to use them. So, there is a need to diagnose glaucoma accurately with low cost. Hence, a new methodology for an automated diagnosis of glaucoma is required. Fundus images are used for the diagnosis of glaucoma. The effect of glaucoma can be reduced if we can predict glaucoma in its early stages. To predict glaucoma in its early stages a deep learning (DL) and machine learning techniques are used. with convolutional neural network and Bayesian networks. Classifiers, such as convolutional neural networks (CNNs) and Bayesian network, can infer a hierarchical representation of images to discriminate between glaucoma and non-glaucoma patterns for diagnostic decisions. These networks are compared with the parameter classification accuracy to predict which network model is best for classifying Fundus images. The features relevant for distinguishing Glaucoma patients is extracted from the fundus image. The Empirical Wavelet Transform (EWT) is employed for extracting features from fundus image . Use of wavelet transform for the feature extraction, which is faster and enables better resolution and high performance for representation and visualization of the abnormality in fundus image than other methods.