Automatic Classification of Glaucomatous Images using Wavelet and Moment Feature

Automation in retinal medical field is highly adopted as quick and precise diagnosis. Computational decision support is easy and affordable system for early diagnosis of glaucoma to prevent the vision loss. In the proposed methodology, glaucoma detection using wavelet and geometric moment features of image texture are presented. Three wavelet filters Daubechies (db3), Symlets (sym3) and Biorthogonal (bior3.3, bior3.5, bior3.7) are used for image decomposition and higher order moments are used for feature computation. The z-score normalization is applied on features before classification. Three classifiers viz. support vector machine (SVM), k-nearest neighbor (KNN) and Error Back-Propagation Training Algorithm (EBPTA) are employed for classification and respective accuracies are calculated. Standard data base RIM-ONE r2 is used for comparison of existing and proposed method. Proposed algorithm provides better accuracy and less computational time than existing algorithm using wavelet and moment features.