ABSTRACT
Diabetic retinopathy is worldwide disease which causes a major threat to sight. It can lead to blindness among aged people. Diabetic retinopathy is mostly found in people suffering from type II diabetes as compared to type I. Diabetic patients are recommended to undergo retinal tests at regular intervals to detect diabetic retinopathy at early stage, as symptoms are difficult to detect until it reaches its severe stage. In this chapter, a hybrid algorithm is proposed by merging the features of two algorithms to cluster large data. The proposed algorithm has the ability to select the efficient features and unique key points while removing redundant values, considering each and every key point having minute variations and providing better accuracy. The diabetic retinopathy data set is used to evaluate the performance of proposed work along with the existing DBSCAN and K-Means.
