图像去噪中MCM方程的改进AOS数值方案
首发时间:2009-06-10
摘要:平均曲率驱动方程(mean curvature motion, MCM)是具有明确几何意义的一种偏微分方程图像处理的模型.本文针对现有的显式方案,半隐式加性算子分裂方案(additional operator splitting method, AOS)在稳定性和CPU时间等方面的不足,提出了的一种改进AOS方案.讨论了改进AOS方案的CPU时间和稳定性.实验结果表明,改进的AOS方案在能达到稳定的且不亚于半隐式AOS方案的图像去噪效果的同时,较显式方案和半隐式AOS方案大大节省CPU和存储量.
关键词: 平均曲率驱动方程(MCM) 半隐式AOS方案 改进AOS方案 图像去噪
For information in English, please click here
An improved AOS scheme of MCM equation for image denosing
Abstract:Mean curvature motion equation is a PDE model for image processing with specific geometry significance. It is usually implemented by explicit scheme and Semi-implicit AOS scheme. The stability of the explicit scheme is severe. Semi-implicit AOS scheme is of high stability, good rotational invariance but reveals a poor computational complexity and memory requirement. In the paper, an improve AOS scheme is proposed to achieve both of a good stability and complexity. Experiment results show that, the improved AOS scheme is at least 3 times more efficient than the widely-used Semi-implicit AOS scheme and achieves its effectiveness for image denosing.
Keywords: mean curvature motion Semi-implicit AOS Scheme improved AOS scheme denoising
论文图表:
引用
No.3302147861512446****
同行评议
共计0人参与
勘误表
图像去噪中MCM方程的改进AOS数值方案
评论
全部评论0/1000