Image inpainting algorithm based on double cross TV

被引：1

作者：

Zhai, Dong-Hai ^{[1
,2
]}

Duan, Wei-Xia ^{[1
]}

Yu, Jiang ^{[1
]}

机构：

[1] School of Information Science and Technology, Southwest Jiaotong University

[2] Engineering School, Tibet University

来源：

Dianzi Keji Daxue Xuebao/Journal of the University of Electronic Science and Technology of China | 2014年 / 43卷 / 03期

关键词：

Double cross TV algorithm; Image inpainting; Inpainting accuracy; Neighborhood information;

D O I：

10.3969/j.issn.1001-0548.2014.03.020

中图分类号：

学科分类号：

摘要：

The various total variation (TV) algorithms currently reconstruct the lost or deteriorated parts of images by related information of damaged pixel and its 4 neighborhood pixels. Their inpainting accuracy is low because of finite related information. The double cross TV algorithm proposed in this paper divides 8 neighborhood pixels into 2 groups, and computes the pixel value of damaged pixel by using 4 neighborhood pixels in each group respectively. Therefore, the weighted mean of these two pixel values is the final inpainting pixel value. The Experiments show an improvement in PSNR and less iteration number compared with the original TV.

引用

页码：432 / 436

页数：4

共 13 条

[11]

Micchelli C.A., Shen L.-X., Xu Y.-S., Proximity algorithms for the L1/TV image denoising model, Advances in Computational Mathematics, 2, 38, pp. 401-426, (2013)

[12]

Yang J.-F., Zhang Y., Yin W.-T., A fast alternating direction method for TV L1-L2 signal reconstruction from partial Fourier data, IEEE J Sel Top Signal Process, 4, pp. 288-297, (2010)

[13]

Li F., Shen C.-M., Liu R.-H., Et al., A fast implementation algorithm of TV inpainting model based on operator splitting method, Computer and Electrical Engineering, 37, pp. 782-788, (2011)

← 1 2 →