Annotation and Retrieval of Retinal Images using Support Vector Machine with Active Learning

In the medical field, many images are acquired using different modalities like X-ray, Computed Tomography and Magnetic Resonance Imaging. Organizing these images and retrieving them is a critical task. In this paper, a novel retinal image retrieval and automatic annotation of ten different retinal diseases is presented and is comprised of three steps namely, feature extraction, automatic annotation and retrieval of similar images. In feature extraction, visual features like texture, color and shape were derived from the images and stored in a database. The retinal images are affected by various diseases and have to be identified easily and quickly for better treatment. For automatic annotation, we have considered ten different retinal diseases and annotated the retinal image with its corresponding disease using Support Vector Machine (SVM) with Active Learning (AL) and discriminative based automatic image annotation. In retrieval of images, Bray Curtis distance measure was used to retrieve the similar images by comparing the feature values of images stored in the database. Our proposed system has achieved a precision of 82% and specificity of 96.3%.