﻿ 基于结构稀疏的图像修复实验设计 The Design of Experiment on Image Inpainting Using Structure Sparisity

Vol.06 No.03(2016), Article ID:17576,8 pages
10.12677/AE.2016.63019

The Design of Experiment on Image Inpainting Using Structure Sparisity

Chen Hu, Wenjing Yu*, Xiaoyu Zhang, Cheng Fang, Zunhan Yang, Ning Chen

Experimentation Center for Electronics and Information Education, East China University of Science and Technology, Shanghai

Received: Apr. 28th, 2016; accepted: May. 12th, 2016; published: May. 19th, 2016

ABSTRACT

The investigation on the sparisity of natural image is the basis of image inpainting using structure sparisity of images. The experiment is designed to deepen students’ understanding about the concept of structure sparisity and the implementation to the image inpainting technology. In the experiment, the conception and implementation of structure sparisity and sparisity expression are introduced to achieve image inpainting by image patch sparisity. The experiments and comparisons with the classical algorithm are performed. In the paper, the process, virtue and shortage of inpainting in different inpainting algorithms are compared, which is helpful for students to understand the process and the meaning of image inpainting.

Keywords:Image Inpainting, Structure Sparisity, Sparisity Expression

1. 引言

2. 结构稀疏算法的原理 [4]

1) 用结构稀疏性计算图像块的优先权 [4] [7] [8]

(1-1)

(a) Selection ofimage blocks (b) Filling image blocks(a) 图像块选择 (b) 图像块填充

Figure 1. The picture direction on the sparsity of image blocks

(1-2)

(1-3)

(1-4)

2) 利用稀疏表达实现图像块的填充 [4] [7] [8]

Xu算法从已知图像样本块中，选取与待修复块最为相似的个样本块，待修复块近似为这些已知图像块的线性加权组合(如图3(b))。

(1-5)

(1-6)

(1-7)

3. 实验设计

MATLAB作为欧美科学界、工程界、教育界广泛使用的科学计算与工程仿真语言已在我国广大工科院校普及。数字信号处理，数字图像处理课程大都引入MATLAB语言作为解题工具及仿真环境。

MATLAB2010或以上版本含图像处理toolbox，

PSNR计算程序，

BSDC图像库中图像

1) 图像的基本操作 [10] - [12]

Imshow()显示图像

imwrite()把图像写入(保存)图像文件中

impixel确定像素颜色值

imcrop()图像的裁剪

imresize()图像比例缩放变换

exp()指数函数

2) 图像修复算法的实现过程

Figure 2. Flowsheet of image inpainting algorithm based on structure sparsity

3) 修复过程的比较

4) 不同修复算法实现及效果比较

5) 实验结果

(a) (b) (c)(d) (e) (f)

Figure 3. Comparison between the process of different inpainting algorithms

(a) Original image (b) Broken image(c) The repair of Criminisi algorithm (d) The repair of Xu algorithm(a) 原图 (b) 破损图 (c) Criminisi算法修复(d) Xu算法修复

Figure 4. Comparison images of the restoration effect after the removal of the figure

(a) Original image (b) Broken image (c) The repair of criminisi algorithm (d) The repair of Xu algorithm(a) 原图 (b) 破损图 (c) Criminisi算法修复 (d) Xu算法修复

Figure 5. Comparison of the result of scratch image inpainting

(2-1)

4. 结语

The Design of Experiment on Image Inpainting Using Structure Sparisity[J]. 教育进展, 2016, 06(03): 120-127. http://dx.doi.org/10.12677/AE.2016.63019

1. 1. Bertalmio, M., Sapiro, G., Caselles, V., et al. (2000) Image Inpainting. Proceedings of ACM SIGGRAPH, New Orleans, 23-28 July 2000, 417-424.

2. 2. Elad, M., Starck, J.L., Querre, P. and Donoho, D.L. (2005) Simultaneous Cartoon and Texture Image Inpainting Using Morphological Component Analysis. Applied and Computational Harmonic Analysis, 19, 340-358. http://dx.doi.org/10.1016/j.acha.2005.03.005

3. 3. Guleryuz, O.G. (2003) Nonlinear Approximation Based Image Recovery Using Adaptive Sparse Reconstructions. Proceedings of 2003 International Conference on Image Processing, Vol. 1, 713-716. http://dx.doi.org/10.1109/icip.2003.1247061

4. 4. Xu, Z.B. and Sun, J. (2010) Image Inpainting by Patch Propagation Using Patch Sparsity. IEEE Transactions on Image Processing, 19, 1153-1165. http://dx.doi.org/10.1109/TIP.2010.2042098

5. 5. Zhang, Q and Li, J.J. (2012) Examplar-Based Image Inpainting Using Color Distribution Analysis. Journal of Information of Science and Engineering, 28, 641-654.

6. 6. Fadili, M.J., Starck, J.L. and Murtagh, F. (2009) Inpainting and Zooming Using Sparse Representations. The Computer Journal, 52, 64-79. http://dx.doi.org/10.1093/comjnl/bxm055

7. 7. 张晴. 基于样本的数字图像修复技术研究[D]: [博士学位论文]. 上海: 华东理工大学.

8. 8. 孙剑. 基于视觉先验的图像处理模型与算法研究[D]: [硕士学位论文]. 西安: 西安交通大学.

9. 9. Criminisi, A., Pérez, P. and Toyama, K. (2004) Region Filling and Object Removal by Exemplar-Based Inpainting. IEEE Transactions on Image Processing, 13, 1200-1212. http://dx.doi.org/10.1109/TIP.2004.833105

10. 10. 张德丰. 数字图像处理(MATLAB版) [M]. 北京: 人民邮电出版社, 2009: 15-30.

11. 11. 王家文. MATLAB7.6图形图像处理[M]. 北京: 国防工业出版社, 2009: 21-32.

12. 12. 张德丰. MATLAB数字图像处理[M]. 北京: 机械工业出版社, 2012: 276-285.

13. 13. Renaud, R., Baroncini, V. and Gusmao, A.A., et al. (2005) Final Report from the Video Quality Experts Group on the Validation of the Objective Models of Video Quality Assessment. http://www.wqeg.org

14. 14. Wong, A. and Orchard, J. (2008) A Nonlocal-Means Approach to exemplar-Based Inpainting. Proceedings of 15th IEEE International Conference on Image Processing, San Diego, 12-15 October 2008, 2600-2603. http://dx.doi.org/10.1109/icip.2008.4712326

15. 15. Wu, J.Y. and Ruan, Q.Q. (2006) Object Removal by Cross Iso-photes Exemplar-Based Inpainting. Proceedings of 18th International Conference on Pattern Recognition, Hong Kong, 810-813.

*通讯作者。