摘要
LBF(Local Binary Fitting)模型利用局部图像信息能够对强度分布不均匀的图像进行分割,然而,该算法仅考虑均值信息,导致模型在处理弱边界图像时得不到理想的分割结果。为此提出一种改进方法:在考虑图像局部均值信息的同时考虑图像局部方差信息和全局方差信息,使得演化曲线能够准确地停止在目标边界上;同时为了加快曲线演化的速度,结合了CV模型的能量项。实验结果表明,改进的方法对含有弱边界信息图像进行分割时能取得较好的效果,演化速度上也有明显的提高。
The LBF（Local Binary Fitting） method can segment image with inhomogeneous intensity using local image information.However,the method can not attain ideal segmentation outcome when dealing with the image with weak boundaries due to its sole reference on mean information.By analysing the phenomenon,an improved method is introduced in this paper,which considers local mean information as well as local and global variance information of the image simultaneously.In this way,the evolution curves can stop at the targeted boundaries accurately.Meanwhile,in order to accelerate the speed of curve evolution,the energy functions of CV model is combined too.A number of experiments prove that the improved method can achieve better effect in segmenting the image with weak boundaries,the evolution speed has been considerably raised as well.
出处
《计算机应用与软件》
CSCD
2011年第2期25-27,33,共4页
Computer Applications and Software
基金
国家自然科学基金(60805003)
关键词
图像分割
LBF模型
弱边界
Image segmentation LBF model Weak boundary
