Screening Diabetic Retinopathy in Developing Countries using Retinal Images
In developing countries, diabetic retinopathy (DR) is the leading cause of blindness in diabetic patients due to intraocular hypertension or high glucose level. Its detection in an earlier stage is essential to prevent vision loss in type 2 diabetic patients. In this paper, the computer aided automatic screening system for diabetic retinopathy is proposed. DR can be diagnosed by detecting the abnormal lesions such as hemorrhages in retinal images and analyzing its relationship with the fovea region. The proposed method consists of the following stages, namely: retinal image enhancement and classification, hemorrhages detection and segmentation, fovea localization and Diabetic Retinopathy classification. The multi directional local histogram equalization is used to enhance the retinal image for better classification rate. The Gabor transform and Support vector machine (SVM) classifier is used for retinal image classifications. The proposed method is tested on publicly available HRFand DIARETDB1datasets. The sensitivity and specificity of hemorrhages detection are 94.76% and 99.85%, respectively. Thus, the severity of Diabetic Retinopathy in Type 2 diabetic patients can be easily identified by detecting fovea region and hemorrhage lesions and analyzing the relation between them to prevent vision loss in diabetic patients.