SVM-BASED TEXTURE CLASSIFICATION IN OPTICAL COHERENCE TOMOGRAPHY

This paper describes a new method for automated texture classification for glaucoma detection using high resolution retinal Optical Coherence Tomography (OCT). OCT is a non-invasive technique that produces cross-sectional imagery of ocular tissue. Here, we exploit information from OCT images, specifically the inner retinal layer thickness and speckle patterns, to detect glaucoma. The proposed method relies on support vector machines (SVM), while principal component analysis (PCA) is also employed to improve classification performance. Results show that texture features can improve classification accuracy over what is achieved using only layer thickness as existing methods currently do.