摘要
提出了一种基于梯度向量流的自动图像分割算法,该算法首先将梯度向量流场转化为一个标量场,该标量场能够显著简化种子点选取和区域增长的步骤。在得到图像的初始分割后,再使用基于区域邻接图的算法来将相似区域合并得到最终分割结果。试验结果表明,该算法能够有效地解决医学图像中多目标区域的自动分割问题。
Image segmentation is an important issue in medical imaging. In this paper, a fully automatic algorithm based on Gradient Vector Flow （GVF） was proposed. Firefly a scalar field was constructed from the GVF field. The scalar field can greatly simplify the selection of seeds and region growing process. Then a Region Adjacency Graph （RAG） based region merging algorithm was applied to get the final result from the initial segmentation. Several experimented results demonstrate that this method is efficient in multiple objects segmentation of medical images.
出处
《计算机应用》
CSCD
北大核心
2007年第1期149-151,共3页
journal of Computer Applications
关键词
医学图像处理
图像分割
梯度向量流
区域邻接图
medical image processing
, image segmentation
GVF（Gradient Vector Flow）
RAG（ Region Adjacency Graph）
