Segmentation of retinal blood vessels using Gabor wavelet and morphological reconstruction

Abstract

Retinal blood vessels analysis from colour fundus images is conducted by ophthalmologists for retinal diseases diagnoses, such as glaucoma, diabetic retinopathy, maculopathy as well as macular oedema. This study applies the combination of 2D Gabor wavelet with morphological reconstruction on fundus images for obtaining segmented retinal blood vessels. The proposed approach is validated on sixty colour retinal images taken from two database, i.e. DRIVE and STARE. The performance of this study is evaluated by measuring sensitivity, specificity and accuracy. On DRIVE database, the accuracy which is obtained by the proposed approach is up to 95.87%. While on STARE database, the accuracy obtained is 89.46%. These results show that the proposed approach has successfully performed in segmenting retinal blood vessels on colour retinal images.