摘要
参数活动轮廓模型(Snakes)分割图像时有两个明显的缺陷:要求初始轮廓线位于图像特征附近,且对深度凹陷区域的分割也不理想。基于梯度向量流(GVF)Snakes较好地解决了传统Snakes的两个本质缺陷。但是,由于GVF Snakes内力的性质和GVF的光滑性,使得对曲率大的边缘点不能精确定位。该文通过采用各向异性方程对图像扩散平滑和边缘增强,改善计算势能力场和梯度向量力场指向边缘的精确度。两种力场的有机组合作为Snakes的外力场。这种新的力场Snakes具有GVF Snakes和势能力场Snakes的优点,对左心室核磁共振图像(MRI)进行分割能得到精确的边缘轮廓。
Parametric active contour model(Snakes) manifests two limitations: an initial contour must be set near the feature of the image and such a model cannot deal properly with the concave regions in the image.By replacing the gradient field with the gradient vector flow(GVF),Snakes can segment concave region edges effectively and has a large capture range.GVF Snakes has poor performances at high curvature boundaries(such as corners) due to the smoothness of the intrinsic force and the gradient vector flow.The image is smoothed by an anisotropic diffusion equation.The combination of the gradient vector force field with the potential force field is proposed for Snakes.Experiments demonstrate that the model curve is driven accurately to the object boundary by the new forces even if the initial estimate curve is not close,the object is nonconvex or the edge has a high local curvature on the left ventricle MRI.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2006年第1期76-80,共5页
Journal of Nanjing University of Science and Technology
关键词
参数活动轮廓
梯度向量流
势能力场
图像分割
parameter active contour
gradient vector flow
potential force field
image segmentation
