摘要
肺血管分割一直是重要而困难的工作,因此,提出了一种新的基于几何形变模型的三维肺血管图像分割方法。现有几何形变模型方法仅仅包含了图像区域与边缘属性中的一种,而新方法能够同时包含上述两种图像属性。首先,定位血管内外同质区域;然后,通过目标边缘能量的计算使曲面沿着图像梯度方向的二阶导数进行演化,以使其精确收敛到目标边缘;最后,根据上述步骤,建立三维血管分割形变模型。通过多组CT图像的实验表明,该方法快速、准确,对背景噪声具有较好的适应性。
Lung vessel segmentation has been an important and difficult task, therefore, a new 3D lung vessel segmentation scheme based on geometric active contour model is proposed in this paper. Existing active contour models only contain image region or edge property, but this scheme can contain both of them. First, inside and outside vessel homogeneous regions are located. Second, in order to converge to vessel edge accurately, on the basis of the computation of target edge energy, surface evolution is made along the second order derivative in the direction of the image gradient. Third, based on the above steps, the active contour model of 3D vessel segmentation is constructed. The experiments on multiple-series of lung CT images show that this scheme is accurate and fast, and has high adaptability to background noise.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2010年第10期2296-2301,共6页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(60671050)
中央高校基本科研业务费专项资金(NO90304001)
东北大学博士后基金(20100519)资助项目
关键词
肺血管分割
形变模型
水平集
边缘检测
CT图像
lung vessel segmentation
active contour model
level set
edge detection
CT image
