摘要
Chan-Vese模型以其能较好地处理图像的模糊边界和复杂拓扑结构而广泛运用于图像分割中。但针对含灰度不均匀性和复杂背景的图像分割效果较差。提出一种基于图割方法的自适应分片常数式的CV模型。首先通过计算距离函数和邻域相似度修正CV模型的拟合项及长度项;然后根据像素点和拟合中心的相似度控制划分拟合中心所在的区域,以备准确估计拟合中心;最后利用图割算法最小化能量函数并得到新的拟合中心以进行下一轮最小化,从而得到更准确且高效的分割结果。
Chan-Vese model,which has better ability to handle the blurry boundary and complex topological structures in images,has been widely used in image segmentations.However,the effect on segmentation in the images with intensity inhomogeneity and multiple-backgrounds isn＇t fine.An adaptive multiple piecewise constant CV model based on graph cut is proposed.First,distance function and local similarities are calculated to revise the optimization term and the length term of CV model.Based on the similarity of the current pixel and the optimization centers,the region of the optimization centers in different parts can be partitioned to estimate the optimization centers accurately.And then the energy function is minimized by graph cut algorithm generating new optimization centers for the next step,which can get the more accurate and efficient segmentation results.
出处
《计算机工程与应用》
CSCD
北大核心
2010年第28期13-16,共4页
Computer Engineering and Applications
基金
国家自然科学基金No.60802039
高等学校博士学科点专项科研基金No.200802880018
国家高技术研究发展计划(863)No.2008AA121103
南京理工大学自主科研项目(No.2010ZYT070)~~
关键词
CV模型
分片常数
自适应
拟合中心
图割
Chan-Vese （CV） model
piecewise constant
adaptive
optimization center
graph cut
