摘要
对医学图像进行有效的去噪并保持边缘信息,有利于图像的后续处理.本文分析P-M模型和Gilboa的复扩散模型以及它们的不足,提出一种改进的各向异性复扩散模型.该方法先用中值滤波对图像进行预处理,去除梯度值大的噪声点,然后用图像的虚部求扩散系数,以此引导扩散模型中的边缘检测函数,再进行八邻域像素的扩散过程.实验表明,该方法能达到较理想的去噪和保持边缘的效果,而且减少了迭代次数,缩短了计算时间.
For the purpose of further processing,it is necessary to remove noisy while keeping the edge＇s information in medical processing. In this paper,the P-M model and Gilboa＇s complex diffusion model as well as their disadvantages are discussed. An improved anisotropic complex diffusion model for medical image denoising is proposed. The method applies median filter to pre-processing the image for removing the pixels with high gradient firstly. Then use the image＇s imaginary part for diffusion coefficient solving,so as to adjust edge detection function of the diffusion model. Finally,diffuse the image with eight-neighborhood pixels. Experiments show that the method can remove noises and keep the edge of the image. Also,it can reduce the iterations and shorten the computing time.
出处
《小型微型计算机系统》
CSCD
北大核心
2010年第5期969-973,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60673063)资助
国家科技支撑计划项目(2007BAH11B02)资助
浙江省自然科(Y1080436)资助
浙江省科技计划项目(2009C32007
2009C31106)资助
关键词
医学图像去噪
偏微分方程
P-M模型
各向异性复扩散
medical image denoising
partial differential equations
P-M model
anisotropic complex diffusion
