一种基于图像平均灰度值的快速图像修复算法

彭坤杨; 董兰芳

doi:10.11834/jig.20100109

图像处理和编码 | 浏览量 : 0 下载量: 654 CSCD: 0

PDF
导出
分享
收藏
专辑

一种基于图像平均灰度值的快速图像修复算法
A Fast Image Inpainting Algorithm Based on Average Gray Value
2010年15卷第1期页码：50
纸质出版：2010
DOI： 10.11834/jig.20100109
稿件说明：

移动端阅览

彭坤杨, 董兰芳. 一种基于图像平均灰度值的快速图像修复算法[J]. 中国图象图形学报, 2010,15(1):50. DOI： 10.11834/jig.20100109.

PENG Kun-yang, DONG Lan-fang. A Fast Image Inpainting Algorithm Based on Average Gray Value[J]. Journal of Image and Graphics, 2010, 15(1): 50. DOI： 10.11834/jig.20100109.

摘要

基于纹理合成的图像修复技术用于修复大面积破损区域，目前此类算法都存在时间复杂度高的缺点。针对纹理算法的匹配技术进行改进，提出了一种基于图像平均灰度值的快速图像匹配算法。该算法在匹配之前预先计算纹理块的平均灰度值以及分割后的纹理块的局部平均灰度值，以取代计算复杂的匹配项SSD（sum of squared differences）；匹配过程只需对平均灰度值进行快速比较，结合阈值控制筛选掉大部分候选纹理块。实验结果表明，该算法在不损害图像修复质量的同时，将纹理修复的效率提高到实时水平。同时在纹理合成和纹理修复

Abstract

Image inpainting techniques based on texture synthesis can be used to remove large objects from digital images. However

these algorithms have high time complexity. To improve the matching technology in inpainting

this paper presents a efficient algorithm that is based on average gray value. Average gray values of texture blocks are calculated to avoid complex calculation of SSD (sum of squared differences). Then in the matching process

most blocks are rapidly eliminated by comparing their average gray values with threshold control. The result shows that the algorithm obtains both real time performance and high quality result. The proposed algorithm has general applicability in texture synthesis and image inpainting.