Keypoint detectors and texture analysis based comprehensive comparison in different color spaces for automatic detection of the optic disc in retinal fundus images

Abstract Detection of the optic disc which has similar brightness with hard and soft exudate lesions seen in early stage diabetic retinopathy is very difficult due to different light conditions contrast values. Automatic detection these by expert systems medical field important. In this context, we propose a new approach based on analysis color spaces, keypoint detectors, texture for retinal fundus images. If information contained within actual region, an important consideration automated disc. This study can be divided into five sections, respectively, image preprocessing, processing, detection, analysis, performance evaluation. The analyses patch images compatible keypoints obtained from Red–Green–Blue (RGB) its channels were carried out. was validated Digital Retinal Images Vessel Extraction public dataset. According results, Local Binary Pattern performed region interest around detected detectors presented good RGB green channel