Detection of Diabetic Retinopathy Using Improvised Fuzzy Contextual Data Clustering Method

Abstract

The study introduces the vigorous fuzzy contextual data K-means clustering technique, an advanced version of the traditional k-means clustering method that incorporates localized information parameters customized for each cluster. A comparative analysis is performed between the robust fuzzy local information k-means clustering and the modified fuzzy C means clustering, which enhances Fuzzy C Means with a median adjustment parameter for diabetic retinopathy detection. The three datasets: IDRiD, DIATREB1, and DIATREB2 fundus images are used in this research. The proposed algorithm achieves a 94.4% accuracy rate. It is designed to efficiently categorize a large volume of retinal images, thereby improving performance and addressing the critical need for prompt and accurate diagnoses in diabetic retinopathy care.