Automatic Severity Grading of Optic Disc Edema Using Appearance-Based Features

Optic disc edema (ODE) is one of the abnormal signs of several ophthalmic diseases, such as papilledema, infection, and tumors. ODE severity grading is essential for providing proper treatments to prevent worse consequences. In this work, we used 15 appearance-based features acquired from the localized optic disc region to grade the stage of the severity. The features are extracted from a gray-level co-occurrence matrix (GLCM), color, intensity line profile, and vessels. The linear support vector machine is used to classify 146 ODE images selected from three datasets. The proposed method yields very promising performance of 78.77%, 82.5%, and 74.24% for accuracy, specificity, and sensitivity, respectively. It significantly outperforms a state-of-the-art method by 17.13%, 16.62%, and 18.5%, respectively.