Computer Aided Design Diagnosis for Glaucoma Detection in Retinal Images by Spatial Fuzzy C Means with Level Set Segmentation

Abstract

By the CAD diagnosis using various clinical parameters such as plate-ratio (commander) optical cup are determined to diagnose glaucoma. Hough converter and circular return view disk fundus image is taken. Level compilation methods and integrated space-based blur package are proposed to analyze the optical cup area from the viewing disk. The experiments were performed using the MATLAB software HRF database using the proposed approach to fundus imaging from images in the hospital. The Linear regression fit is intent to discover the Gold standard assessment for the evaluating attained CDR based on the fundus image acquired from the hospital database. Them CDR values is obtained through Bayesian classifies to train the dataset. Consequences formed from the sorting achieve data results, the sensitivity of 96.47%, specificity of 92.85% and an accuracy of 94.83%. Receiver operating characteristic curve is plotted for the observed and gold standard values of CDR. With this approach, the boundaries of the region can be accurately identified and the target mass of screening retinal images for early detection of glaucoma can be used and the resulting segmentation in consistent areas can be made firmer.