当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术2019年18期

一种改进的Criminisi 新算法的图像复原技术
陈美玲,朱铝芬,张云,石瑶
(南京工业大学浦江学院 机电学院,江苏 南京 211134)

摘  要:本文针对Criminisi 算法及其现阶段存在的不足,提出一种基于图像样本块的图像复原算法。通过在Criminisi 算法中加入模块的自适应算子,使原来的9*9 模块在适当情况下自适应变化为其余模块大小,来降低修复误差。实验结果在主观上修复部分更加清晰自然,修复边缘没有太大的瑕疵;从客观评价指标上分析:图像峰值信噪比PSNR 变大,图像模糊程度变小;修复运行时间更长,图像修复效果更好。结果表明改进后的Criminisi 算法从主观层面和客观层面都具有更好的图像修复效果,即污染区域图像断层更少,过渡更加平滑,纹理丰富、结构复杂区域的修复也更加自然。


关键词:自适应;Criminisi 算法;图像修复



中图分类号:TP317.4          文献标识码:A         文章编号:2096-4706(2019)18-0043-04


An Improved Criminisi Algorithm for Image Restoration

CHEN Meiling,ZHU Lyufen,ZHANG Yun,SHI Yao

(Institute of Electrical and Mechanical,Nanjing Tech University Pujiang Institute,Nanjing 211134,China)

Abstract:To overcome the shortcomings of Criminisi algorithm and its present stage,a new improved image restoration algorithm based on sample blocks is proposed. In Criminisi algorithm,the adaptive operator of the module is added to reduce the repair error by adaptively changing the size of the other modules from the original 9*9 module in appropriate cases. The experimental results show that the repair part is more clear and natural in subjective analysis,and the repair edge is not too defective. From the objective evaluation index,the peak signal-to-noise ratio (PSNR) of the image becomes larger,and the image blurring degree becomes smaller. The repair operation time is longer,and the image restoration effect is better. The experimental results show that the improved Criminisi algorithm has a better image restoration effect both in subjective and objective aspects,that is,the contaminated area has fewer image slices,more smoother,and the texture-rich and complex areas are more natural.

Keywords:adaptive operator;Criminisi algorithm;image restoration


课题项目:南京工业大学浦江学院2018 年校级科研重点项目:基于双目立体视觉的目标测距技术研究(项目编号:njpj2018-1-01)。


参考文献:

[1] GUILLEMOT C,LE M O. Image Inpainting:Overview and Recent Advances [J].IEEE Signal Processing Magazine,2014,31(1):127-144.

[2] 张斐. 大区域图像修复算法研究 [D]. 南京:南京邮电大学,2013.

[3] BERTALMIO M,SAPIRO G,CASELLES V,et al. Image Inpainting [C]//Proceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques. Proceedings of the SIGGRAPH.New York:ACM Press/Addison-Wesley Publishing Co.,2000:417-424.

[4] CRIMINISI A,PEREZ P,TOYAMA K. Region filling and object removal by examplar-based inpainting [J].IEEE Transactions on Image Processing,2004,13(9):1200-1212.

[5] 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):50-55.

[6] 沈峘,李舜酩,毛建国,等. 数字图像复原技术综述 [J].中国图象图形学报,2009,14(9):1764-1775.

[7] 孟春芝,何凯,焦青兰. 自适应样本块大小的图像修复方法 [J]. 中国图象图形学报,2012,17(3):337-341.

[8] WEXLER Y,SHECHTMAN E,IRANI M. Space-Time Completion of Video [J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2007,29(3):463-476.


作者简介:陈美玲(1984-),女,满族,吉林长春人,讲师,硕士,研究方向:控制工程,图像处理。