Fast Detection and Segmentation in Retinal Blood Vessels using Gabor Filters

Abstract

The relationship between change in retinal vessels morphology and progress to detect the signs of diabetic retinopathy in the early stages has been the major subject of ophthalmologists. However, the accurate segmentation and fast detection of vessels to diagnosis disease have been given important role in quality of application. In this paper, in order to capture high frequency information a bank of 180 Gabor filters is used. Additionally, a systematic way of determining the threshold value for reliable performance is described. The proposed approach is applied to images from DRIVE retina database. The results of efficient segmentation indicate that time of detection is considerable. The accuracy of the segmentation is better and the time required to perform the segmentation is faster than many other methods in the literature.