摘要
提出了一种新搜索策略下的快速图像修复算法。通过定义新的优先权计算函数,克服了图像低纹理区域修复过于滞后的问题。又通过预测修复后图像块统计属性,对所有待匹配的图像块进行筛选,加快了图像修复速度,改善了修复效果。实验结果显示,该算法适用于多种类型的数字图像修复。
This paper proposed a quick image inpainting algorithm under a new searching strategy. By a new priority computation function, it overcame the problem caused by the hysteresis of the low texture regions inpainting, and by forecasting the statistical property of the image patch to be inpainted, selected the source image patches which accelerated the inpainting process, improved the inpainting results. Experiments confirm that this algorithm is suitable to various digital images inpainting.
出处
《计算机应用研究》
CSCD
北大核心
2009年第5期1991-1993,1996,共4页
Application Research of Computers
关键词
图像修复
搜索策略
优先权
匹配函数
images inpainting
searching strategy
priority
matching function