|
福州大学学报(自然科学版) 2015
基于改进C-V模型乳腺癌MR图像分割
|
Abstract:
:在乳腺癌MR图像分割中,传统C-V模型没有充分利用图像边界曲率信息,需要重新初始化水平集函数使其保持为一个符号距离函数(SDF),导致图像分割比较慢,同时目标区域易产生过度分割. 为此,通过在传统的C-V模型中引入惩罚能量项和全局边界曲率能量项,提出一种改进的C-V模型图像分割方法,克服了水平集函数需要重新初始化和目标区域易产生过度分割等问题. 实验表明,改进的C-V模型对乳腺癌MR图像具有较好的分割效果,分割收敛速度较快.
n segmentation of MR images of breast cancer,traditional C-V model does not make full use of the image boundary and curvature information,needing to re-initialize the level set function to keep it as a signed distance function (SDF). It results in slower image segmentation. It is very easy that target area is over-segmentation. In this paper,we introduce the penalties energy and global boundary curvature energy in the traditional C-V model and propose a improved C-V model image segmentation method. This new model overcomes the shortages that the level set function needs to be re-initialize and the target area is easy to produce over-segmentation and other issues. The experimental results show that the improved C-V model has a better segmentation results o breast cancer in MR images