Automated segmentation and quantitative analysis of optic disc and fovea in fundus images

Abstract Fundus image is widely used diagnosis method and involves the retinal tissues which can be important biomarkers for diagnosing diseases. Many studies have proposed automatic algorithms to detect optic disc (OD) fovea. However, they showed some limitations. Although precise regions of are clinically important, most these focused on localization not segmentation. Also, did sufficiently prove clinical effectiveness methods using quantitative analysis. Furthermore, many them researched about single tissue. To compensate limitations, this study automated segmentation both OD In study, dataset was acquired from DRIVE Drions databases, additional ground truth obtained an ophthalmologist. The original fundus preprocessed remove noise enhance contrast. And vessel segmented use fovea step, a region interest designated based features increase accuracy. segment OD, removed substituted intensity value four nearest non-vessel pixels. Finally, were including intensity, shape size. evaluated by analysis eight methods. As result, high performance with accuracy 99.18 99.80 % database.