摘要
在人类视觉系统特性基础上提出了一个改进的各向异性平滑方法。加入图像四阶偏微分信息避免"阶梯效应",扩展演化方向,改进数值模型。实验结果表明,该方法在噪声消除和边缘保留方面能获得较好的效果。
An improved anisotropic smoothing method based on the characteristics of human visual system is proposed.The four-order partial differential information of image is merged into this method to avoid the＂blocky effects＂ ,the direction of diffusion is increased and the numerical model is improved.Experimental results show that this method can obtain good resuits in noise elimination and edge reservation.
出处
《计算机工程与应用》
CSCD
北大核心
2010年第26期155-157,共3页
Computer Engineering and Applications
基金
国家自然科学基金No.10671156~~
关键词
各向异性扩散
人类视觉系统
图像平滑
anisotropic diffusion
human visual system
image smoothing
