Identification of Diabetic Damage Detection in Retinal Images using ANFIS

Abstract

Digital retinal images plays very crucial role inmedical domain and applications. The role of these images
vary across different diseases like micro aneurysms, hemorrhages, hard exudates, cotton wool spots or venous circles
etc. One of the important application of retinal image processing is required in diagnosing and treatment of many
diseases affecting the retina and the choroid behind. With huge volume of images being taken from the drive database, it
is imperative to have an automated classifier system. In this work Adaptive Neuro Fuzzy Inference System (ANFIS) has
been designed and developed for detection of abnormalities in the retinal images. Using wavelet decomposition &
Principal Component Analysis (PCA), image features are extracted & segregated. These image features form the basis
for ANFIS based classification system. The proposed approach has been validated with help of data sets comprising over
40 images sourced from DRIVE data base. The classifier performance has been evaluated through different performance
measures and this research demonstrates the suitability of proposed approach in identifying retinal diseases.