Blood Vessel Extraction Method Based On Color Fundus Images

In clinical diagnosis and medical research,blood vessels are very important biological tissues.Blood vessels in the eye are the only deeper microvascular which can be directly non-invasive observed.As an important feature of fundus images,the morphological changes in the structure are closely related to the disease degree,severity and post-treatment of cardiovascular diseases such as hypertension,diabetes and arteriosclerosis.Therefore,the effective segmentation and extraction of a complete retinal vascular network from the fundus image has important practical significance for medical research and clinical applications.This article first analyzes the research significance of retinal vascular network extraction in fundus images and the research status at home and abroad,and then focuses on the retinal blood vessel segmentation method based on color fundus image.The main work is as follows:1)Aiming at the characteristics of fundus retinal vessels of blurred image detail,uneven illumination of the fundus image,obvious noise and low contrast,the fundus image is uniformly pre-processed to eliminate the influence of these factors as much as possible,so as to improve the quality of the fundus image.2)The advantages and disadvantages of existing vascular network segmentation methods are analyzed.Aiming at the application requirements of general fundus image segmentation and the development trend of medical image segmentation algorithms,a blood vessel segmentation algorithm combining maximum entropy threshold and Otsu method is proposed.3)In order to verify the segmentation effect of the proposed method,the performance of the proposed algorithm was tested by using the fundus image data of the confessed DRIVE and STARE library,and compared with the similar retinal vessel segmentation methods.The experimental results show that the segmentation method used in this paper is more complete and accurate for the segmentation of tiny blood vessels and vascular end points,and has improved the segmentation accuracy.In addition,for the fundus image with lesions,the algorithm can basically segment and extract the vascular network,which proves the feasibility and effectiveness of the proposed algorithm.