Automatic eye screening method based on a multi-anatomical retinal system segmentation with the help of morphology & fuzzy logic

Abstract

Eye examination is used to determine Retinal problems at an early stage. Retina screening can be used to detect various hidden retinal problems including pre-diabetes and diabetes. The doctors will depend on segmentation results of retinal structures to determine the abnormalities in the eye. A novel algorithm for instant identification and segmentation of the optic disk (OD) and detection of exudates in human retinal images is described in this document. The proposed technique uses segmentation methodology for automatically detecting of these structures. This methodology uses morphology theories and fuzzy techniques to detect the structures even in pathological images. The performance is evaluated using two datasets that is DRISHI-GS (optic disc) and DIARETD1 (exudates). Present methods have accuracy and efficiency is less, but in the proposed new novel method i.e morphology based fuzzy logic, we achieved an accuracy 92.91% and PPV (positive predictive value) is 89.54% in optic disc. specificity and F-score also trained with help of subjected method 99% and 93.17% in exudate.