Development of a screening tool for staging of diabetic retinopathy in fundus images

Abstract

Diabetic retinopathy is a condition of the eye of diabetic patients where the retina is damaged because of long-term diabetes. The condition deteriorates towards irreversible blindness in extreme cases of diabetic retinopathy. Hence, early detection of diabetic retinopathy is important to prevent blindness. Regular screening of fundus images of diabetic patients could be helpful in preventing blindness caused by diabetic retinopathy. In this paper, we propose techniques for staging of diabetic retinopathy in fundus images using several shape and texture features computed from detected microaneurysms, exudates, and hemorrhages. The classification accuracy is reported in terms of the area (A z ) under the receiver operating characteristic curve using 200 fundus images from the MESSIDOR database. The value of A z for classifying normal images versus mild, moderate, and severe nonproliferative diabetic retinopathy (NPDR) is 0.9106. The value of A z for classification of mild NPDR versus moderate and severe NPDR is 0.8372. The A z value for classification of moderate NPDR and severe NPDR is 0.9750.

IDRiD: Diabetic Retinopathy – Segmentation and Grading Challenge

“…Similarly, Antal and Hajdu (2014) proposed a strategy involving image-level quality assessment, pre-screening followed by lesion and anatomical features extraction to finally decide about the presence of DR using ensemble of classifiers. Further, for identification of different stages of DR features from morphological region properties (Yun et al, 2008), texture parameters (Acharya et al, 2012;Mookiah et al, 2013b), non-linear features of the higher-order spectra Acharya et al (2008), hybrid Dhara et al (2015) and information fusion (Niemeijer et al, 2009) approaches were found useful. As the DME is graded based on the location of the EXs from macula, many researchers (Giancardo et al, 2012;Medhi and Dandapat, 2014;Perdomo et al, 2016;Marin et al, 2018) proposed EXs based features to determine the severity of the DME.…”