Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization

Jize Xue; Yongqiang Zhao

doi:10.1109/CMVIT.2017.22

Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization

Jize Xue, Yongqiang Zhao

School of Automation

Northwestern Polytechnical University Xian

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

15 Scopus citations

Abstract

In hyperspectral imagery denoising, rank-1 tensor decomposition (R1TD) model can utilize the spatial and spectral information jointly and reduce the noise efficiently. It is difficult to estimate the rank of hyperspectral imagery accurately, and the rank uncertainty will make the R1TD denoising algorithm inefficient. The nonlocal similar patches have lower rank than image, it can be used in rank-1 tensor decomposition process instead of explicitly estimating rank parameters. In this work, a nonlocal low-rank regularization is introduced to avoid the rank uncertainty to influence denoising performance. Then an alternating direction method of multipliers (ADMM) optimization technique is designed to solve the minimum problem. Compared with the state of art methods, proposed algorithm significantly improves the hyperspectral imagery quality both in visual inspection and image quality indices.

Original language	English
Title of host publication	Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	40-45
Number of pages	6
ISBN (Electronic)	9781509049936
DOIs	https://doi.org/10.1109/CMVIT.2017.22
State	Published - 14 Mar 2017
Event	2017 International Conference on Machine Vision and Information Technology, CMVIT 2017 - Singapore, Singapore Duration: 17 Feb 2017 → 19 Feb 2017

Publication series

Name	Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017

Conference

Conference	2017 International Conference on Machine Vision and Information Technology, CMVIT 2017
Country/Territory	Singapore
City	Singapore
Period	17/02/17 → 19/02/17

Keywords

hyperspectral image denoising
low-rank regularization
nonlocal patches grouping (NPG)
rank estimation bias
rank-1 tensor decomposition (R1TD)

Access to Document

10.1109/CMVIT.2017.22

Cite this

Xue, J., & Zhao, Y. (2017). Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization. In Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017 (pp. 40-45). Article 7878652 (Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CMVIT.2017.22

Xue, Jize ; Zhao, Yongqiang. / Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization. Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 40-45 (Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017).

@inproceedings{9e4f2f37d1ea4e2fac8257a2bd1296d4,

title = "Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization",

abstract = "In hyperspectral imagery denoising, rank-1 tensor decomposition (R1TD) model can utilize the spatial and spectral information jointly and reduce the noise efficiently. It is difficult to estimate the rank of hyperspectral imagery accurately, and the rank uncertainty will make the R1TD denoising algorithm inefficient. The nonlocal similar patches have lower rank than image, it can be used in rank-1 tensor decomposition process instead of explicitly estimating rank parameters. In this work, a nonlocal low-rank regularization is introduced to avoid the rank uncertainty to influence denoising performance. Then an alternating direction method of multipliers (ADMM) optimization technique is designed to solve the minimum problem. Compared with the state of art methods, proposed algorithm significantly improves the hyperspectral imagery quality both in visual inspection and image quality indices.",

keywords = "hyperspectral image denoising, low-rank regularization, nonlocal patches grouping (NPG), rank estimation bias, rank-1 tensor decomposition (R1TD)",

author = "Jize Xue and Yongqiang Zhao",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.; 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017 ; Conference date: 17-02-2017 Through 19-02-2017",

year = "2017",

month = mar,

day = "14",

doi = "10.1109/CMVIT.2017.22",

language = "英语",

series = "Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "40--45",

booktitle = "Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017",

}

Xue, J & Zhao, Y 2017, Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization. in Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017., 7878652, Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017, Institute of Electrical and Electronics Engineers Inc., pp. 40-45, 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017, Singapore, Singapore, 17/02/17. https://doi.org/10.1109/CMVIT.2017.22

Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization. / Xue, Jize; Zhao, Yongqiang.
Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 40-45 7878652 (Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017).

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

TY - GEN

T1 - Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization

AU - Xue, Jize

AU - Zhao, Yongqiang

PY - 2017/3/14

Y1 - 2017/3/14

N2 - In hyperspectral imagery denoising, rank-1 tensor decomposition (R1TD) model can utilize the spatial and spectral information jointly and reduce the noise efficiently. It is difficult to estimate the rank of hyperspectral imagery accurately, and the rank uncertainty will make the R1TD denoising algorithm inefficient. The nonlocal similar patches have lower rank than image, it can be used in rank-1 tensor decomposition process instead of explicitly estimating rank parameters. In this work, a nonlocal low-rank regularization is introduced to avoid the rank uncertainty to influence denoising performance. Then an alternating direction method of multipliers (ADMM) optimization technique is designed to solve the minimum problem. Compared with the state of art methods, proposed algorithm significantly improves the hyperspectral imagery quality both in visual inspection and image quality indices.

AB - In hyperspectral imagery denoising, rank-1 tensor decomposition (R1TD) model can utilize the spatial and spectral information jointly and reduce the noise efficiently. It is difficult to estimate the rank of hyperspectral imagery accurately, and the rank uncertainty will make the R1TD denoising algorithm inefficient. The nonlocal similar patches have lower rank than image, it can be used in rank-1 tensor decomposition process instead of explicitly estimating rank parameters. In this work, a nonlocal low-rank regularization is introduced to avoid the rank uncertainty to influence denoising performance. Then an alternating direction method of multipliers (ADMM) optimization technique is designed to solve the minimum problem. Compared with the state of art methods, proposed algorithm significantly improves the hyperspectral imagery quality both in visual inspection and image quality indices.

KW - hyperspectral image denoising

KW - low-rank regularization

KW - nonlocal patches grouping (NPG)

KW - rank estimation bias

KW - rank-1 tensor decomposition (R1TD)

UR - http://www.scopus.com/inward/record.url?scp=85017264991&partnerID=8YFLogxK

U2 - 10.1109/CMVIT.2017.22

DO - 10.1109/CMVIT.2017.22

M3 - 会议稿件

AN - SCOPUS:85017264991

T3 - Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017

SP - 40

EP - 45

BT - Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017

Y2 - 17 February 2017 through 19 February 2017

ER -

Xue J, Zhao Y. Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization. In Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 40-45. 7878652. (Proceedings - 2017 International Conference on Machine Vision and Information Technology, CMVIT 2017). doi: 10.1109/CMVIT.2017.22

Rank-1 tensor decomposition for hyperspectral image denoising with nonlocal low-rank regularization

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this