Automatic Segmentation of Blood Vessels in Colour Retinal Images using Spatial Gabor Filter and Multiscale Analysis
Automatic Segmentation of Blood Vessels in Colour Retinal Images using Spatial Gabor Filter and Multiscale Analysis
复制标题
使用空间 Gabor 滤波器和多尺度分析自动分割彩色视网膜图像中的血管
DOI：
10.1007/978-3-540-92841-6_66
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
K. G. Prabhu
中科院分区：
文献类型：
--
作者：
P. Siddalingaswamy;K. G. Prabhu
视网膜血管是眼科图像中重要的解剖结构。血管的自动分割是计算机辅助诊断系统中的重要步骤之一，用于检测影响人类视网膜的疾病，如糖尿病视网膜病变。我们提出了一种使用空间Gabor滤波器分割视网膜血管的方法，因为它们可以被调谐到特定的频率和方向。定义了一个新的参数，以便于在不同尺度下进行滤波，以检测不同厚度的血管。该方法进行了测试的40彩色视网膜图像的DRIVE（数字视网膜图像血管提取）数据库与手动分割地面真理。视网膜血管分割的总体准确率为84.22%。
Retinal blood vessels are significant anatomical structures in ophthalmic images. Automatic segmentation of blood vessels is one of the important steps in computer aided diagnosis system for the detection of diseases such as Diabetic Retinopathy that affect human retina. We propose a method for the segmentation of retinal blood vessels using Spatial Gabor filters as they can be tuned to the specific frequency and orientation. A new parameter is defined to facilitate filtering at different scales to detect the vessels of varying thicknesses. The method is tested on forty colour retinal images of DRIVE (Digital Retinal Images for Vessel Extraction) database with manual segmentations as ground truth. An overall accuracy of 84.22% is achieved for segmentation of retinal blood vessels.