CAJ | 학술논문

这篇文章用一种新方法对眼底图血管中心线进行提取.这种方法用双边滤波而不是通常用的高斯滤波来去除图像噪声.这种方法包括3个步骤,首先是用双边滤波改进的canny算子来提取出所有的血管;然后,用检测到的血管来获得其相应血管交叉部分,应用局部极大和极小抑制方法就可以得到相应血管样本侧面;最后,用高斯模型匹配血管样本侧面,高斯匹配的二阶微分零交叉就代表血管的边界位置,高斯匹配的峰值就代表血管的中线位置.这种方法输出亚像素精度的血管中心线.
저편문장용일충신방법대안저도혈관중심선진행제취.저충방법용쌍변려파이불시통상용적고사려파래거제도상조성.저충방법포괄3개보취,수선시용쌍변려파개진적canny산자래제취출소유적혈관;연후,용검측도적혈관래획득기상응혈관교차부분,응용국부겁대화겁소억제방법취가이득도상응혈관양본측면;최후,용고사모형필배혈관양본측면,고사필배적이계미분령교차취대표혈관적변계위치,고사필배적봉치취대표혈관적중선위치.저충방법수출아상소정도적혈관중심선.
This paper presents a new method to track the center line of retinal blood vessel with sub-pixel accuracy. The method is based on a modified Canny edge detection method with a bilateral filter. In this method, Rather than Gaussian filter as it normally used, a bilateral filter is used in Canny edge detection method to remove the image noise and provide a better vessel cross-section profiles. There are three stages of the method. In the first stage, the modified Canny detection with bilateral filter is used to detection all edges of the vessels. In the second stage, the edges of the vessels are then used to obtain their corresponding vessel crossing-sections. And the vessel sample profiles are searched from the vessel crossing-sections using local maximum and local minimum. In the last stage, the Gaussian model is used to fit the vessel sample profiles, and the zeros cross of the second order derivatives of the Gaussian fit vessel sample profiles are then used to represent the boundaries of the blood vessels. The peaks of the Gaussian fits are then used as positions of vessel center lines. The method outputs the center lines in sub-pixel accuracy.