Retinal Image Enhancement and Eye Disease Identification

Abstract

Digital image processing has left an engraved influence in almost every technical field such as in engineering, medical etc. Early detection is vital in many vision losing diseases like cataract, diabetic retinopathy, diabetic mellitus cataract which cause blindness among working people even in younger ages. Retina will give all vital information about eye. Therefore retinal image analysis has become one major disease diagnosis method in modern ophthalmology. The main purpose of this study is enhancement of color fundus imageries centered on luminosity and contrast and to detect eye diseases such as diabetic retinopathy, cataract, and diabetic mellitus cataract. Due to complex imaging environment images are usually under uneven illumination, blurred and low contrast condition, which is not advantageous for identification, mainly in automated image examining systems. Thus this paper proposes an image enhancement technique to ameliorate color fundus image brightness (luminosity) as well as contrast. A luminance gain matrix is used to enhance the luminosity of RGB components whereas contrast, the main factor is improved by luminosity component of Lab color scheme using CLAHE. Once enhancement has been completed then detection of eye diseases such as diabetic retinopathy, cataract and diabetic mellitus cataract performed. Early recognition of these protects the people from loss of vision which result in blindness. Thus, this project proposes an image processing technique for enhancement of retinal imageries and to detect the eye diseases in advance.