基于块结构稀疏度的自适应图像修复算法

doi:10.3969/j.issn.0372-2112.2013.03.022

PDF(3133 KB)

电子学报 ›› 2013, Vol. 41 ›› Issue (3) : 549-554. DOI: 10.3969/j.issn.0372-2112.2013.03.022

学术论文

基于块结构稀疏度的自适应图像修复算法

李志丹¹, 和红杰¹, 尹忠科¹, 陈帆¹, 仁青诺布²

作者信息 +

Adaptive Image Inpainting Algorithm Based on Patch Structure Sparsity

LI Zhi-dan¹, HE Hong-jie¹, YIN Zhong-ke¹, CHEN Fan¹, RENQING Nuo-bu²

Author information +

文章历史 +

摘要

现有基于稀疏性的图像修复算法采用固定大小的待填充块和邻域一致性约束,且在全局搜索待填充块的最优匹配块,既降低了待修复区域的结构连贯性和纹理清晰性,又增加了算法的时间复杂度.针对上述问题,根据破损区域特性和块结构稀疏度间的关系,提出基于块结构稀疏度的自适应图像修复算法.根据最大优先权值点的块结构稀疏度值,设定不同参数以自适应选取待填充块大小、邻域一致性约束权重系数和局部搜索区域大小,并通过仿真实验分析讨论了各参数选取.实验结果表明本文算法较文献算法在峰值信噪比上提高0.3dB~1.2dB,并且提高算法速度3~7倍.

Abstract

In the existing patch sparsity based image inpainting algorithms,the exemplar-size and neighborhood-consistence weight are fixed,and the best match patches of the patch to be filled are searched in the whole source region.However,it decreases the connectivity of structure and clearness of texture while increases the time complexity of this algorithm.To address these problems,an adaptive image inpainting algorithm is proposed based on patch structure sparsity,in the light of the relationship between the characteristics of damage region and patch structure sparsity.According to patch structure sparsity value of the point which has the maximal prority value,the size of patch to be filled,the neighborhood consistence weight and the part-search region size are adaptively confirmed through setting some parameters,then these parameters are analysed and discussed by some experiments.Experimental results show that the PSNR is increased by 0.3~1.2dB and the speed is improved by 3~7 times compared with the existing algorithms.

导出引用

李志丹, 和红杰, 尹忠科, 陈帆, 仁青诺布. 基于块结构稀疏度的自适应图像修复算法[J]. 电子学报, 2013, 41(3): 549-554. https://doi.org/10.3969/j.issn.0372-2112.2013.03.022

LI Zhi-dan, HE Hong-jie, YIN Zhong-ke, CHEN Fan, RENQING Nuo-bu. Adaptive Image Inpainting Algorithm Based on Patch Structure Sparsity[J]. Acta Electronica Sinica, 2013, 41(3): 549-554. https://doi.org/10.3969/j.issn.0372-2112.2013.03.022

中图分类号： TP391.41

参考文献

[1] Bertalmio M,Sapiro G,Caselles V,et al.Image inpainting [A].Proceedings of ACM SIGGRAPH [C].New Orleans:ACM Press,2000.417-424.
[2] Chan T,Shen J.Mathematical models for local nontexture inpaintings[J].SIAM Journal on Applied Mathematics,2001,62(3):1019-1043.
[3] Criminisi A,Perez P,Toyama K.Region filling and object removal by exemplar-based image inpainting[J].IEEE Transactions on Image Processing,2004,13(9):1200-1212.
[4] 朱为,李国辉.基于自动结构延伸的图像修补方法[J].自动化学报,2009,35(8):1041-1047. ZHU Wei,LI Guo-Hui.Image completion based on automatic structure propagation.Acta Automatica Sinica,2009,35(8):1041-1047.(in Chinese)
[5] 雷鸣,王春东,等.一种新的样本块图像修补方法[J].光电子·激光,2009,20 (5):677-689. LEI Ming,WANG Chun-dong,et al.A new exemplar-based image completing method.Journal of Optoelectronics · Laser,2009,20(5):677-689.(in Chinese)
[6] 张岩,孙正兴,姚伟.基于方向经验模型分解的图像修复方法[J].电子学报,2010,38 (2):257-262. ZHANG Yan,SUN Zheng-xing,YAO Wei.Image completion based on direction empirical mode decomposition[J].Acta Electronica Sinica,2010,38(2):257-262.(in Chinese)
[7] Wong A,J Orchard.A nonlocal-means approach to exemplar-based inpainting [A].IEEE International Conference on Image Processing [C].San Diego,CA,USA:IEEE Press,2008.2600-2603.
[8] Bugeau A,Bertalmio M,Caselles V,et al.A comprehensive framework for image inpainting[J].IEEE Transactions on Image Processing,2010,19(10):2634-2645.
[9] Wu J Y,Ruan Q Q,An G H.Exemplar-based image completion model employing PDE corrections[J].Informatica,2010,21(2):259-276.
[10] 蔡泽民,赖剑煌.一种基于超完备字典学习的图像去噪方法[J].电子学报,2009,37(2):347-350. CAI Ze-min,LAI Jian-huang.An over-complete learned dictionary-based image de-noising method[J].Acta Electronica Sinica,2009,37(2):347-350.(in Chinese)
[11] 孙玉宝,韦志辉,等.多形态稀疏性正则化的图像超分辨率算法[J].电子学报,2010,38(12):2898-2902. SUN Yu-bao,WEI Zhi-hui,et al.Multimorphology sparsity regularized image super-resolution[J].Acta Electronica Sinica,2010,38(12):2898-2902.(in Chinese)
[12] Shen B,Hu W,Zhang Y M,et al.Image inpainting via sparse representation [A].IEEE International Conference on Acoustics,Speech and Signal Processing [C].Taipei,Taiwan:IEEE Press,2009.697-700.
[13] Wang Y X,Zhang Y J.Image inpainting via weighted sparse non-negative matrix factorization [A].IEEE International Conference on Image Processing [C].Brussels,Belgium:IEEE Press,2011.3409-3412.
[14] Xu Z B and Sun J,Image inpainting by patch propagation using patch sparsity[J].IEEE Transactions on Image Processing,2010,19(5):1153-1165.
[15] Wang Z,Bovik A C,Sheikh H R,Image quality assessment:from error visibility to structural similarity[J].IEEE Transactions on Image Processing,2004,13(4):600-612.