Classification of Diabetic Macular Edema and Its Stages Using Color Fundus Image  

作　　者：Muhammad Zubair Shoab A.Khan Ubaid Ullah Yasin 

机构地区：[1]the Department of Computer Engineering,National University of Sciences & Technology,Islamabad 44000,Pakistan [2]Army Medical College & Armed Forces Institute of Ophthalmology,National University of Sciences & Technology,Islamabad 44000,Pakistan

出　　处：《Journal of Electronic Science and Technology》2014年第2期187-190,共4页电子科技学刊（英文版）

摘　　要：Diabetic macular edema （DME） is a retinal thickening involving the center of the macula. It is one of the serious eye diseases which affects the central vision and can lead to partial or even complete visual loss. The only cure is timely diagnosis, prevention, and treatment of the disease. This paper presents an automated system for the diagnosis and classification of DME using color fundus image. In the proposed technique, first the optic disc is removed by applying some preprocessing steps. The preprocessed image is then passed through a classifier for segmentation of the image to detect exudates. The classifier uses dynamic thresholding technique by using some input parameters of the image. The stage classification is done on the basis of anearly treatment diabetic retinopathy study （ETDRS） given criteria to assess the severity of disease. The proposed technique gives a sensitivity, specificity, and accuracy of 98.27%, 96.58%, and 96.54%, respectively on publically available database.Diabetic macular edema （DME） is a retinal thickening involving the center of the macula. It is one of the serious eye diseases which affects the central vision and can lead to partial or even complete visual loss. The only cure is timely diagnosis, prevention, and treatment of the disease. This paper presents an automated system for the diagnosis and classification of DME using color fundus image. In the proposed technique, first the optic disc is removed by applying some preprocessing steps. The preprocessed image is then passed through a classifier for segmentation of the image to detect exudates. The classifier uses dynamic thresholding technique by using some input parameters of the image. The stage classification is done on the basis of anearly treatment diabetic retinopathy study （ETDRS） given criteria to assess the severity of disease. The proposed technique gives a sensitivity, specificity, and accuracy of 98.27%, 96.58%, and 96.54%, respectively on publically available database.

关 键 词：Disk diameter EDEMA FOVEA optic disk hard exudates MACULA RETINA 

分 类 号：R587.2[医药卫生—内分泌] R774.5[医药卫生—内科学]