Diagnosis of diabetic retinopathy and thyroid using retinal image

Diabetic retinopathy(DR) is one of the complication related with diabetes and also a major cause of blindness in middle and older age groups. The spatial analysis of microaneurysms, exudates and hemorrhages can help to determine the severity of DR. The development of computer assisted retinal image analysis technology may help the clinicians to reduce their time of screening; and mass screening is also possible. In this preliminary work, initially preprocessing was carried out by noise removal. Then watershed segmentation followed by active contour segmentation was implemented to segment the blood vessel regions and some region properties of blood vessels have been extracted to diagnose into different abnormalities such as hemorrhages, soft and hard exudates and small red dots. The algorithm was tested using the retinal images collected from Kaggle and 91 % of diagnostic accuracy was achieved.