Automated Glaucoma Detection Using Variational Mode Decomposition from Fundus Images

Abstract

Glaucoma is a chronic eye disorder and one of the major causes of vision loss. Increased intraocular pressure damaged the optic nurves and hence blindness. Available methods on glaucoma image classification are expensive and slow. Therefore fast and low cost methods are needed. In this paper, glaucoma image classification using two dimensional variational mode decomposition and support vector machine from fundus images is proposed. The variational mode decomposition is used to decompose the glaucoma and normal images. Features are extracted from decomposed sub band images. Selected and reduced features are used to classify images in glaucoma or normal by support vector machine. The obtained accuracy, sensitivity, specificity are 94.17 %, 95 %, and 95 %, respectively for tenfold cross validation technique. Obtained results confirm that proposed method is adequate and improved over the state-of-the-art methods.