A review on texture-based methods for anomaly detection in retinal optical coherence tomography images

Abstract

Retina is an important body organ responsible for human vision. There are many important retinal diseases which may damage the vision and even cause blindness. Optical Coherence Tomography (OCT) is an important tool for verifying and evaluating retina. Moreover, retina is a window to brain and since the imaging from retina is simpler than brain, valuable information from brain can be obtained through retinal imaging. In fact, many neuro-degenerative diseases can be followed through retinal OCT images. Texture refers to the way of locating pixels with different intensity values in an image neighborhood. In this paper, the purpose is to provide a review on the methods which focus on the anomaly detection in OCT images. Also, the role of texture in the identification of retinal diseases is discussed in detail. Different texture descriptors in image processing applications are introduced and the methods which utilized them for the mentioned purpose are explained. In addition, different retinal diseases are classified in several classes and the texture-based methods suggested for each class of diseases are separately described. © 2023 Elsevier GmbH