Retinal Image Analysis Using Machine Learning

The morphology or structure of veins in retinal pictures is a significant marker of maladies like glaucoma, diabetic retinopathy and hypertension. The precision of veins in retinal image division influences the nature of retinal picture investigation which is utilized in analysis strategies in current ophthalmology. Differentiation upgrade is one of the urgent strides in any of retinal vein division draws near. This paper exhibits an evaluation of the appropriateness of an as of late created spatially versatile difference improvement strategy for upgrading retinal fundus pictures for vein division. The upgrade system was functioned with a slight variation in Tyler Coye calculation, which is upgraded with Hough line change based vessel recreation strategy. The proposed system is assessed on the two open and accessible dataset which are STARE and DRIVE. The retinal blood vessels are extracted using Tyler Coye algorithm. After segmentation, the extracted blood vessel is given as input to different types of machine learning algorithm for the detection of diabetic retinopathy disease. Thus, the proposed pipeline is useful to prevent the disease that cause permanent blindness at early stages, so that the people can take necessary precautions to cure the diabetic retinopathy, before they get affected by the permanent blindness.