彩色眼底图像视网膜血管分割方法研究进展

A Survey of Retinal Vessel Segmentation in Fundus Images

视网膜血管分割方法是眼科计算机辅助诊断和大规模疾病筛查系统的基础,文中讨论了基于彩色眼底图像的视网膜血管分割方法研究进展.概述了该领域的背景意义、常用标准库、性能衡量指标、采用的主要算法及其优缺点,旨在快速地引导研究人员了解本领域研究内容.视网膜血管分割方法可分为基于血管跟踪的方法、基于匹配滤波的方法、基于形态学处理的方法、基于形变模型的方法和基于机器学习的方法等5大类,各类方法都各有特点,为后期研究提供了基础.其中基于机器学习的方法是目前最重要的分割方法,以数据驱动的方式为眼科辅助诊断系统提供依据.尽管研究人员已经做了大量工作,视网膜血管分割依然有进一步提高精度和效率的空间.眼底图中其他生理结构和各种病灶的干扰,微小血管、视盘内血管、新生毛细血管网等的分割,都是血管分割问题中有待解决的难点.

Retinal vessel segmentation is the basis of the ophthalmic disease computer-aided diagnosis and large-scale screening system. This paper reviews the progress of retinal vessel segmentation in fundus image. Paper outlines the background and significance of this research, the commonly used standard databases, performance metrics, the advantages and disadvantages of the vessel segmentation algorithms. It is aimed at quickly guiding researchers to understand the contents of this field. The method of retinal vessel segmenta-tion can be divided into five main categories: blood vessel tracking, matched filtering, mathematical mor-phology, deformable model based, and machine learning. All the methods contain their own characteristics and contribute to the latter researches, among which machine learning based method is the most important one. It provides the decision support for the computer-aided diagnosis with clues by data-driven approach. Although researchers have done a lot of work, retinal vessel segmentation still can be improved in accuracy and efficiency. There are many difficulties to be resolved in retinal vessel segmentation, such as the inter-ference by physiological structure and lesions, and the segmentation of microvascular, vessels on the optic disc and intraretinal microvascular abnormalities (IRMA).