| 求同存异的非局部图像去噪 |
| |
| 引用本文： | 张选德,冯象初,王卫卫,魏立力.求同存异的非局部图像去噪[J].中国科学:信息科学,2013(7):907-919. |
| |
| 作者姓名： | 张选德 冯象初 王卫卫 魏立力 |
| |
| 作者单位： | [1]西安电子科技大学应用数学系,阳安710071 [2]宁夏大学数学计算机学院信息系,银川750021 |
| |
| 基金项目： | 国家自然科学基金（批准号：61001156,61105011,60872138,61271294,11261044）; 宁夏大学自然科学基金（批准号：ZR1206）资助项目 |
| |
| 摘 要： | 非局部方法是近年来图像恢复领域最重要的方法之一,这类方法的关键在于如何理解和描述自然图像的自相似性质.本文通过对自相似性质的深入分析,提出描述自相似性质的两个原则：1）＂两方向＂原则：利用隐含在图像中的两方向相关结构.2）＂求同存异＂原则：利用相似性的同时保持各相似块之间的相对差异.然后利用奇异值分解建立了一种符合这两个原则的非局部图像去噪模型.该模型能获得很好的去噪效果.
|
| 关 键 词： | 图像去噪 自相似 两方向相关结构 求同存异 |
Exploit the similarity while preserving the difference for nonlocal image denoising |
| |
| ZHANG XuanDe,FENG XiangChu,.,WANG WeiWeiI,z WEI LiLi.Exploit the similarity while preserving the difference for nonlocal image denoising[J].Scientia Sinica Techologica,2013(7):907-919. |
| |
| Authors: | ZHANG XuanDe FENG XiangChu WANG WeiWeiI z WEI LiLi |
| |
| Institution: | 1 Department of Applied Mathematics, School of Science, Xidian University, Xi＇an 710071, China; 2 School of Mathematics and Computer, Ningxia University, Yingchuan 750021, China |
| |
| Abstract: | Nonlocal method is one of the most important methods in the field of image restoration in recent years, the nutshell of this method is the interpretation and the modeling of the self-similarity property （SSP） of natural images. In this paper, we perform a deep analysis of the SSP and accordingly propose two principles for nonlocal image modeling （1. Exploiting the two direction correlation structures inherent in natural images, 2. Exploiting the similarity and simultaneously preserving the difference between similar patches）. On the basis of these principles, we develop a new nonlocal model for image denoising by using singular value decomposition. Numerical experiments indicate that the proposed model leads to competitive denoising results. |
| |
| Keywords: | image denoising self-similarity two-direction correlation structure exploit the similarity while preserving the difference |
| 本文献已被 维普 等数据库收录！ |
|
