基于相似度滤波算法的视网膜血管分割

在计算机辅助眼底图像视网膜血管分割中,基于匹配滤波算法的应用非常广泛.而传统匹配滤波器算法存在分割细小血管效果较差、噪声多以及视盘干扰等问题.本文提出一种相似度滤波算法的眼底图像视网膜血管分割方法.首先用多层阈值和水平集算法提取视盘干扰区域,利用高斯模糊去除视盘干扰区域.然后采用相似度滤波运算对去除视盘干扰的彩色眼底图像进行处理.最后,将余弦相似度图进行二值化后与余弦相似度加强图进行区域连通性判断,实现眼底图像视网膜血管分割.结果 表明,该算法能较好地分割细小血管以及去除视盘干扰,能更为准确地提取眼底图像视网膜血管.

RETINAL VASCULAR SEGMENTATION BASED ON SIMILARITY FILTERING ALGORITHM

In computer-aided retinal vascular segmentation of fundus images, matched filter algorithm is widely used. However, the traditional matched filter algorithm has some problems such as poor performance and optic disc interference. A similarity filter algorithm for retinal vascular segmentation in fundus images is proposed in this paper. Firstly, multi-level threshold and level set algorithm are used to extract optic disc interference regions, and Gaussian blur is used to remove optic disc interference region. Then the similarity filter algorithm is used to process the color fundus images without optic disc interference. Finally, the cosine similarity image is binarized and the cosine similarity enhancement image is combined to achieve retinal vascular segmentation in fundus images. The results show that the algorithm can better segment the small vessels and remove the interference of optic disc, and can extract retinal vessels in fundus images more accurately.