动静脉血管自动分类方法及其管径测量

发布时间：2018-05-21 16:28

本文选题：眼底图像 + 血管特征向量　；参考：《仪器仪表学报》2017年09期

【摘要】：视网膜动静脉管径以及动静脉比值可以反映高血压患者脑卒中发病的风险,因此对视网膜血管直径的量化分析有助于病情的风险评估和防治工作。提出了一种视网膜动静脉自动分类和血管直径的自动测量方法。首先,对视网膜血管网络进行分割,并获取中心线;其次选取了不同颜色空间中的不同通道分量,提出了基于中心线像素和血管像素的特征向量、血管宽度和中心光反射的特征向量,采用K均值聚类实现感兴趣测量区域内动静脉的自动分类;最后统计血管横截面的灰度曲线分布,利用高斯曲线进行拟合,根据半高度全宽获取动静脉宽度。分别对REVIEW和DRIVE数据库进行实验,验证了本方法的有效性。
[Abstract]:Retinal arteriovenous diameter and arteriovenous ratio can reflect the risk of stroke in hypertensive patients, so quantitative analysis of retinal vascular diameter is helpful to risk assessment and prevention of the disease. A method for automatic classification of retinal arteriovenous vessels and automatic measurement of vessel diameters is presented. Firstly, the retinal vascular network is segmented and the center line is obtained. Secondly, different channel components in different color spaces are selected, and the feature vectors based on the center line pixels and vascular pixels are proposed. The feature vector of vascular width and central light reflection is used to realize the automatic classification of arteries and veins in the area of interest by K-means clustering. Finally, the grayscale curve distribution of the cross section of blood vessel is calculated and fitted by Gao Si curve. The arterial and venous widths are obtained according to the full width of the half height. Experiments on REVIEW and DRIVE databases are carried out to verify the effectiveness of this method.
【作者单位】：福州大学物理与信息工程学院;福建农林大学计算机与信息学院;
【基金】：福建省自然科学面上基金(2016J01297) 福建省中青年教师教育科研项目(JAT160398,JAT160070)资助
【分类号】：R318;TP391.41

【相似文献】