Advanced Glaucoma Detection Using Hybrid Approach and Singular Value Decomposition from Fundus Images

Abstract

Glaucoma is a known to be an of eye disease that is accountable for impaired vision worldwide. For medical images, image decomposition has high impact for diagnosis of diseases. This work proposes an intelligent hybrid approach-based image decomposition method for advanced glaucoma detection. The proposed approach employs Discrete Wavelet Transform (DWT) hybridisation with Empirical Wavelet Transform (EWT). Texture features are used and extracted from decomposed image components as these captures subtle variation. To discover effective features, extracted features are added to Singular Value Decomposition (SVD). In addition, efficient features are added to the Support Vector Machine (SVM) algorithm and 95.83 percent accuracy, 91.67 percent sensitivity and 100 percent specificity has been achieved for twelve-fold cross validation. Hence, the approach proposed demonstrates superior efficiency of decomposition.