计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (9): 159-162.DOI: 10.3778/j.issn.1002-8331.2010.09.045

• 图形、图像、模式识别 • 上一篇    下一篇

高阶图像扩散模型的中值公式

李媛媛1,潘振宽1,魏伟波1,朱东伟2,张 楠1   

  1. 1.青岛大学 信息工程学院,山东 青岛 266071
    2.亚信科技(中国)有限公司,北京100086
  • 收稿日期:2009-09-24 修回日期:2009-12-29 出版日期:2010-03-21 发布日期:2010-03-21
  • 通讯作者: 李媛媛

Higher order image diffusion model based on median formula

LI Yuan-yuan1,PAN Zhen-kuan1,WEI Wei-bo1,ZHU Dong-wei2,ZHANG Nan1   

  1. 1.College of Information Engineering,Qingdao University,Qingdao,Shandong 266071,China
    2.AsiaInfo Technologies(China),Inc.,Beijing 100086,China
  • Received:2009-09-24 Revised:2009-12-29 Online:2010-03-21 Published:2010-03-21
  • Contact: LI Yuan-yuan

摘要: 经典的TV(Total Variation)模型在对图像扩散的同时能有效保持图像边缘,但该类模型所得到的结果具有明显的阶梯效应,其改进的方案之一是在能量泛函中增加高阶项。但其对应的偏微分方程计算效率非常低。基于中值公式开展了如下研究:给出了TV-L2TV-L1变分图像扩散模型中值公式的四邻域、八邻域计算过程实现;提出了基于散度的高阶图像扩散的中值公式。实验证明高阶TV模型能很好地消除阶梯效应,将中值公式应用于图像扩散模型,提高了计算效率。

关键词: 变分方法, 图像扩散, 中值公式, TV模型

Abstract: The classic TV(Total Variation) model can be used to diffuse images with edge preserving,but it also can result in staircase effects beyond edges.One of the means to reduce staircase effects is to add high order item.But the computation efficiency is low which results in considerable attention in image processing area.Some improvements and applications are presented,which include:Implementations of TV-L2TV-L1 models are given using four points neighborhood and eight points neighborhood.A new median formula is proposed for higher order image diffusion based on divergence to reduce staircase effects.The median formula is used to complex computation of multiscale denoising model to improve computing efficiency.

Key words: variational method, image diffusion, median formulas, TV model

中图分类号: