The method for diabetic retinopathy diagnosis using image processing algorithm

摘要

The invention identifies diabetic retinopathy, a characteristic retinal circulation disorder that occurs in patients with diabetes, by applying image processing technology based on the image acquired by the fundus camera and detecting intraocular hard exudate and retinal hemorrhage. An object of the present invention is to provide a method for diagnosing diabetic retinopathy using an image processing algorithm characterized by the method. Based on the fundus photo acquired by the fundus camera, through image analysis applying an image processing programming technique, retinal hemorrhage caused by microvessels weakened by diabetes or fat components in the blood leak out of the blood vessels and accumulate on the retina. By accurately detecting hard exudate and displaying it to the user, the purpose is to recognize it in the early stage of the onset and prevent the rapid deterioration of the condition. Therefore, in the present invention, since the R component is included as the largest proportion among the color distributions for extracting features in composing the fundus image due to the characteristics of the fundus image acquired by the fundus camera, gray for extracting features by selecting the R component By converting the image into an image, image processing algorithms such as image erosion or image dilatation are applied for the purpose of clearly distinguishing the average brightness value constituting the image and the brightness value indicated by hard exudate or retinal hemorrhage to accurately determine the feature points by the threshold To provide a scientific method for determining whether diabetic retinopathy by extracting and confirming the coordinates.