Abstract

Retinal image segmentation plays an important role in monitoring and diagnosing retinal diseases. It is considered as one of the most challenging tasks for computer graphics researchers. Many researchers used complex approach to obtain the segmentation of retinal images for medical diagnoses. However, the main goal of this study was to propose an alternative method for segmenting retinal image by employing two tools. ShaderMap and GIMP, the GNU Image Manipulation Program used in this paper for creating normal map that can improve the retinal image and give an alternative of those acquired from segmentation. The performance of the proposed method was evaluated on one high resolution retinal image datasets (REVIEW) and one low resolution image datasets (DRIVE). According to the pre-designed questionnaire, the proposed method improved the retinal image and GIMP gave better result compared to ShaderMap. However, when high resolution retinal image used there was no significant differences between both tools.