The diabetic retinopathy illness spreads polygenic disease on the membrane vessels therefore they lose blood provide that causes cavity in brief time, thus early detection of polygenic disease prevents visual defect in additional than five hundredth of cases. The retinal image identification is a very important methodology for diabetic retinopathy detection and analysis. During this paper the formula improved median filter is employed for pre-processing and additional feature extraction of exudates is completed by improved fuzzy bunch formula. The projected system consists of 4 stages. First is the gathering of real time retinal pictures from the hospitals. Second stage is pre-processing of retinal image exploitation improved median filtering. Third stage is feature extraction of Exudates and fourth is usually recommended for corresponding treatment for additional use. Simulation exploitation mat work were done employing a set of pictures and are established that the formula holds smart for all the photographs, exudates may be detected effectively.


Median filter, exudates, Structure pictures, improved fuzzy clustering, proper treatment.