Automatic Optic Disc Localization and Optic Cup Segmentation from Monocular Color Retinal Images for Glaucoma Assessment

摘要

In recent years, computer-aided screening and diagnosis of glaucoma have made considerable progress. Computer-aided retinal image analysis provides an instant detection of retinal features before specialist inspection. Eye disease of glaucoma is a significant globally is due to an increase in intraocular pressure. It causes permanent vision loss, and also early prognosis is complicated. The present work provides an image processing technique used to automatically detect the center of the optic disc and segment the optic cup for prediction of glaucoma by developing software algorithms in MATLAB. Three kinds of steps are used such as pre-processing, detection of optic disc center and optic cup. In the first step, green Chanel image and filtering methods used for removing noises. In the second step, Region localization using entropy is applied to locate Optic Disc center. In the third step, the Optic Cup has segmented by region growing technique. This subject to different retinal fundus image datasets such as DRIONS and DRIVE. Proposed algorithm is obtained 100% Accuracy rate(A), 98% Sensitivity(S), 99% Specificity(Sp), 95% Precision(P), 87% F-score(Fs), 94% G-mean(Gm) and trivial computation time.