摘要
针对噪声图像,基于曲线演化理论与水平集方法,提出一个对噪声鲁棒的水平集分割模型。利用图像局部和全局信息,构造一个新的速度函数,得到一个水平集演化偏微分方程。实验表明,该模型对含有高噪声的合成和真实图像有很好的分割效果,同时能准确提取弱边缘和模糊边缘,而且对轮廓初始化有很强的鲁棒性。
For noisy images,based on curve evolution theory and level set method,in this paper we propose a level set segmentation model robust to noise. It uses local and global image information to construct a new speed function,and derives evolution partial differential equation for level set. Experimental results show that the proposed model has desirable segmentation performance for synthetic and real images with high level noises,meanwhile it can accurately extract weak and fuzzy edges. In addition,this model is very robust to initialisation of contours.
出处
《计算机应用与软件》
CSCD
2016年第4期173-176,195,共5页
Computer Applications and Software
基金
国家自然科学基金项目(11371384)
关键词
图像分割
噪声图像
水平集方法
偏微分方程
Image segmentation
Noisy image
Level set method
Partial differential equation
