LAUSR

The importance of the research is to detect disorders in retinal images using machine learning techniques. Eyes play an important role in our day to day life and are the most valuable gift we have in retinal eye research, the retinal vessel parameters and accurate AVR measurement is consider as the important issues in image processing techniques. Diseases like glaucoma, exudates, and diabetic retinopathy can be observed in earlier stages by using retinal images. Diabetic retinopathy also known as diabetic eye disease which affects up to 80% of all patients who have diabetes eventually leads to blindness. To diagnosis the various disorder in earlier stage by using the retinal image based on the methodology of image processing and machine learning techniques. The retinal image is used to detect the diabetes in early stages by evaluating all the retinal blood vessels together. The proposed novel algorithm known as multi-resolution curvelet Transform and normalized graph cut segmentation to detect the optic disc and blood vessels in the fundus images efficiently. In earlier stage of this research, the pre-processing of fundus image operation for image filtration and color contrast enhancement, and next is image segmentation for blood vessels achieved by image processing techniques such as thresholding, texture, and morphological operation and finally machine learning classification algorithm are executed using convolutional neural network. This construction results in a multi resolution, local, and directional image expansion using contour segments. Results indicate that neural network is better than the other techniques for vessels classification.