Retinal Vessel Segmentation Based on Multi-scale Frangi Filter YUAN Pan,CHEN Yimum-147
(School of Electronic Engineering and Automation,Guilin University of Electronic Technology,Guilin 541004,China)
Abstract:Aiming at the problem that the fundus retinal image has low contrast and i
s disturbed by the boundary of the diseased area,it is difficult to correctly extract the details of the blood vessel. A method of retinal blood vessel segmentation based on multi-scale Frangi filter is proposed. Firstly,the image is preprocessed,and then multi-scale operations are performed on the basis of the Frangi filter to complete the enhancement of image details;finally,Otsu optimized by genetic algorithm is used for threshold segmentation to get the final result graph. Using the proposed method to conduct experiments on the DRIVE database,the simulation results show that the above method has a good effect on the extraction of small blood vessels and has strong practical value.
Keywords:image processing;retinal vessel;genetic algorithm;threshold segmentation
0 引 言
中国是世界上糖尿病患者最多的国家,且逐年上涨,糖尿病会引起视网膜血管产生变化,因此研究视网膜血管分割对于糖尿病的诊断起着积极的作用。本人单位和医院合作共同研究视网膜血管分割技术,视网膜血管分割技术有利于促进智能医疗的发展[1]。然而视
和嫩太阳能庭院灯网膜血管分割容易受到光照、噪声、病变的影响,难以精确分割出血管。
针对上述问题,文献[2]提出一种基于Frangi滤波器和形态学重建的视网膜血管分割方法,首先运用Frangi滤波器检测出血管,再使用形态学重建技术获得清晰的视网膜血管图像。该方法可以简单分割出血管,但是噪声、病变对其影响较大。文献[3]提出一种基于Frangi滤波器和Otsu视网膜血管分割的方法,此方法在文献[2]的基础上融合形态学阈值分割的图像,从而得到最终分割结果图,该方法能有效解决光照、噪声、病变的影响,较为完整的分割出视网膜血管图像,但血管细节难以很好的分割出来。
为应对上述方法的不足,本文提出一种基于多尺度Frangi滤波器的视网膜血管分割方法。首先对彩眼底图像预处理,得到更加有利于后续操作需要的图像;其次在Frangi滤波器的基础上多尺度增强图像细节,有效提升血管与背景的对比度,最后使用遗传算法优化的Otsu进行阈值分割,得到细节完整的血管图像。
1 基于多尺度Frangi滤波器视网膜血管分割
1.1 图像预处理
丰乳贴
彩眼底视网膜图像有红、绿、蓝三个通道。我们对彩眼底视网膜图像通道分离,其中绿通道眼底图像有对比度较高、信噪比低、血管脉络清晰的特点,故本文选取绿通道图像进行后续的处理[4]。