Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior

L. Chen; Hongzhi Zhang; Dongwei Ren; David Zhang; Wangmeng Zuo

doi:10.1007/978-3-662-45643-9_2

Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior

L. Chen^*
, Hongzhi Zhang
, Dongwei Ren
, David Zhang
, Wangmeng Zuo

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Scopus citations

Abstract

As a fundamental tool, L₀gradient smoothing has found a flurry of applications. Inspired by the progress of research on hyper-Laplacian prior, we propose a novel model, corresponding to L_p-norm of gradients, for image smoothing, which can better maintain the general structure, whereas diminishing insignificant texture and impulse noise-like highlights. Algorithmically, we use augmented Lagrangian method (ALM) to efficiently solve the optimization problem. Thanks to the fast convergence rate of ALM, the speed of the proposed method is much faster than the L₀gradient method. We apply the proposed method to natural image smoothing, cartoon artifacts removal, and tongue image segmentation, and the experimental results validate the performance of the proposed algorithm.

Original language	English
Title of host publication	Pattern Recognition - 6th Chinese Conference, CCPR 2014, Proceedings
Editors	Shutao Li, Yaonan Wang, Chenglin Liu
Publisher	Springer Verlag
Pages	12-21
Number of pages	10
ISBN (Electronic)	9783662456422
DOIs	https://doi.org/10.1007/978-3-662-45643-9_2
State	Published - 2014
Externally published	Yes
Event	6th Chinese Conference on Pattern Recognition, CCPR 2014 - Changsha, China Duration: 17 Nov 2014 → 19 Nov 2014

Publication series

Name	Communications in Computer and Information Science
Volume	484
ISSN (Print)	1865-0929
ISSN (Electronic)	1865-0937

Conference

Conference	6th Chinese Conference on Pattern Recognition, CCPR 2014
Country/Territory	China
City	Changsha
Period	17/11/14 → 19/11/14

Keywords

Augmented lagrangian method
Hyper-Laplacian gradient prior
Image smoothing

Access to Document

10.1007/978-3-662-45643-9_2

Cite this

Chen, L., Zhang, H., Ren, D., Zhang, D., & Zuo, W. (2014). Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior. In S. Li, Y. Wang, & C. Liu (Eds.), Pattern Recognition - 6th Chinese Conference, CCPR 2014, Proceedings (pp. 12-21). (Communications in Computer and Information Science; Vol. 484). Springer Verlag. https://doi.org/10.1007/978-3-662-45643-9_2

@inproceedings{a7629fa0b1084361ad0a109cc8439c41,

title = "Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior",

abstract = "As a fundamental tool, L0gradient smoothing has found a flurry of applications. Inspired by the progress of research on hyper-Laplacian prior, we propose a novel model, corresponding to Lp-norm of gradients, for image smoothing, which can better maintain the general structure, whereas diminishing insignificant texture and impulse noise-like highlights. Algorithmically, we use augmented Lagrangian method (ALM) to efficiently solve the optimization problem. Thanks to the fast convergence rate of ALM, the speed of the proposed method is much faster than the L0gradient method. We apply the proposed method to natural image smoothing, cartoon artifacts removal, and tongue image segmentation, and the experimental results validate the performance of the proposed algorithm.",

keywords = "Augmented lagrangian method, Hyper-Laplacian gradient prior, Image smoothing",

author = "L. Chen and Hongzhi Zhang and Dongwei Ren and David Zhang and Wangmeng Zuo",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2014.; 6th Chinese Conference on Pattern Recognition, CCPR 2014 ; Conference date: 17-11-2014 Through 19-11-2014",

year = "2014",

doi = "10.1007/978-3-662-45643-9\_2",

language = "英语",

series = "Communications in Computer and Information Science",

publisher = "Springer Verlag",

pages = "12--21",

editor = "Shutao Li and Yaonan Wang and Chenglin Liu",

booktitle = "Pattern Recognition - 6th Chinese Conference, CCPR 2014, Proceedings",

address = "德国",

}

Chen, L, Zhang, H, Ren, D, Zhang, D & Zuo, W 2014, Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior. in S Li, Y Wang & C Liu (eds), Pattern Recognition - 6th Chinese Conference, CCPR 2014, Proceedings. Communications in Computer and Information Science, vol. 484, Springer Verlag, pp. 12-21, 6th Chinese Conference on Pattern Recognition, CCPR 2014, Changsha, China, 17/11/14. https://doi.org/10.1007/978-3-662-45643-9_2

Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior. / Chen, L.; Zhang, Hongzhi; Ren, Dongwei et al.
Pattern Recognition - 6th Chinese Conference, CCPR 2014, Proceedings. ed. / Shutao Li; Yaonan Wang; Chenglin Liu. Springer Verlag, 2014. p. 12-21 (Communications in Computer and Information Science; Vol. 484).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior

AU - Chen, L.

AU - Zhang, Hongzhi

AU - Ren, Dongwei

AU - Zhang, David

AU - Zuo, Wangmeng

N1 - Publisher Copyright: © Springer-Verlag Berlin Heidelberg 2014.

PY - 2014

Y1 - 2014

N2 - As a fundamental tool, L0gradient smoothing has found a flurry of applications. Inspired by the progress of research on hyper-Laplacian prior, we propose a novel model, corresponding to Lp-norm of gradients, for image smoothing, which can better maintain the general structure, whereas diminishing insignificant texture and impulse noise-like highlights. Algorithmically, we use augmented Lagrangian method (ALM) to efficiently solve the optimization problem. Thanks to the fast convergence rate of ALM, the speed of the proposed method is much faster than the L0gradient method. We apply the proposed method to natural image smoothing, cartoon artifacts removal, and tongue image segmentation, and the experimental results validate the performance of the proposed algorithm.

AB - As a fundamental tool, L0gradient smoothing has found a flurry of applications. Inspired by the progress of research on hyper-Laplacian prior, we propose a novel model, corresponding to Lp-norm of gradients, for image smoothing, which can better maintain the general structure, whereas diminishing insignificant texture and impulse noise-like highlights. Algorithmically, we use augmented Lagrangian method (ALM) to efficiently solve the optimization problem. Thanks to the fast convergence rate of ALM, the speed of the proposed method is much faster than the L0gradient method. We apply the proposed method to natural image smoothing, cartoon artifacts removal, and tongue image segmentation, and the experimental results validate the performance of the proposed algorithm.

KW - Augmented lagrangian method

KW - Hyper-Laplacian gradient prior

KW - Image smoothing

UR - https://www.scopus.com/pages/publications/84914819851

U2 - 10.1007/978-3-662-45643-9_2

DO - 10.1007/978-3-662-45643-9_2

M3 - 会议稿件

AN - SCOPUS:84914819851

T3 - Communications in Computer and Information Science

SP - 12

EP - 21

BT - Pattern Recognition - 6th Chinese Conference, CCPR 2014, Proceedings

A2 - Li, Shutao

A2 - Wang, Yaonan

A2 - Liu, Chenglin

PB - Springer Verlag

T2 - 6th Chinese Conference on Pattern Recognition, CCPR 2014

Y2 - 17 November 2014 through 19 November 2014

ER -

Fast augmented lagrangian method for image smoothing with hyper-Laplacian gradient prior

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this