A Comparative Study of Different Blood Vessel Detection on Retinal Images

Abstract

Detection of blood vessel plays an important stage in different medical areas, such as ophthalmology, oncology, neurosurgery, and laryngology. The significance of the vessel analysis was helped by the continuous overview in clinical studies of new medical technologies intended for improving the visualization of vessels. In this paper, several local segmentation techniques which include such as Vascular Tree Extraction, Tyler L. Coye and Line tracking, Kirsch’s Template and Fuzzy C Mean methods were studied. The main objective is to determine the best approaches in order to detect the blood vessel on the degraded retinal input image (DRIVE dataset). A few Image Quality Assessment (IQA) was obtained to prove the effectiveness of each detection methods. Overall, the result of sensitivity highest came from Kirsch Templates (96.928), while specificity from Fuzzy C means (77.573). However, in term of accuracy average, the Line Tracking method is more successful compared to the other methods.