Automatic Detection of Retinal Vessels using Kirsch's Templates and Region Growing

The retina is the only part in the human body where the blood vessels can be seen directly in vivo and examined for pathological changes. Blood vessel morphology can be an important indicator for many diseases such as diabetic retinopathy, stroke, hypertension, arteriosclerosis, cardiovascular diseases. Analysis of vascular changes can be used in clinical studies, patient screening, and diagnosing ocular diseases. Computer based analysis for automatic assessment of blood vessel anomalies of retinal image initially requires the segmentation of the vessels from background. This work presents an efficient method for automatic detection of blood vessels in retinal images. Firstly green channel of the Retinal image is extracted, since it presents higher contrast between vessel and background. Contrast-limited adaptive histogram equalisation is used to enhance the vessels. Then segmentation is carried out using Kirsch’s compass kernel and Region growing. Morphological closing is applied to close the holes or empty area within the blood vessel created through the Kirsch’s template matching. Finally Median filtering is used to remove the noise. The proposed method has been tested on a set of retinal images. The retinal images were collected from the DRIVE and STARE database. The performance of the proposed algorithm is measured by comparing the obtained output with manually segmented ground truth images.