摘要
图像修复的方法有很多种,目前最常用的有基于偏微分方程的方法和基于纹理合成的方法。这些修复方法运算复杂,运行速度慢。提出了一种快速有效的新方法,从修复区域的边界上选取一个优先级高的点,计算该点与该点邻域信息之间的相似度,利用邻域各点的相似度来估算待修复点的像素值。最后通过实验对结果加以分析,证明了算法是快速有效的。
There are many approaches for image restoration,the most commonly used are based on partial differential equations or texture synthesis method.These algorithms are complicated and time-consuming.This paper presents a new method of quick and efficient.A point is chosen whoes priority is highest on the border of damaged region.According to neighborhood information of this point,the pixel＇s value is calculated to use all the neighborhood similarity. Using this algorithms,estimate all the points in the damaged region.Experiments show that this method is quick and effective.
出处
《计算机工程与应用》
CSCD
北大核心
2010年第7期190-191,220,共3页
Computer Engineering and Applications
基金
安徽省高校自然科学研究计划项目No.2006KJ028B~~
关键词
图像处理
图像修复
邻域信息
优先级
相似度
image processing
image inpainting
neighborhood information
priority
similarity
