Retinal blood vessel segmentation and bifurcation point detection

Abstract

The analysis of retinal blood vessel structure plays an important role in diagnosis of different diseases. Automated extraction of vascular network and identification of bifurcation points can be an important part of computer assisted analysis of retinal vascular disorders. In this paper, we propose an efficient method of automatic blood vessel extraction and bifurcation point detection from retinal images. The proposed method introduces the novel concept of relaxed digital arc for the removal of optic disc region to improve the correctness of the results. Experimental results show the effectiveness of the proposed method. We re-validate the quality of the proposed blood vessel segmentation approach by comparing the segmentation accuracy with existing approaches. The efficiency of bifurcation point detection process is evaluated by comparing manual bifurcation point count with the findings of the proposed approach.