Detection of exudates from RGB fundus images using 3σ control method

Diabetic Retinopathy (DR) is characterized by color based pathologies. The fundus camera generates fundus image in RGB color and ophthalmologist analyze that image and mark the pathologies to diagnose level of DR. Early detection of exudates from retinal images can potentially reduces the risk of blindness of people suffering from DR. The aim of the work presented in this paper is to detect the exudates (yellowish fat deposit on the retinal surface) from fundus images in RGB color space, thereby facilitate a realistic diagnosis close to the method adopted by ophthalmologist. A Statistical measure-three sigma is used to compute the color intensity range of exudates pixels. The retinal images are preprocessed to enhance the color intensity and optic disk (OD) is eliminated because, it shares similar features with exudates. The aim and objective of this paper is to detect the exudates from RGB fundus images, the pre-processed images are then classified based on the information extracted from three-sigma control method. The results so obtained are promising and also facilitates the ophthalmologist in diagnosing the disease.