Tensor-based low-rank graph with multimanifold regularization for dimensionality reduction of hyperspectral images

Jinliang An; Xiangrong Zhang; Huiyu Zhou; Licheng Jiao

doi:10.1109/TGRS.2018.2835514

Tensor-based low-rank graph with multimanifold regularization for dimensionality reduction of hyperspectral images

Jinliang An, Xiangrong Zhang^*, Huiyu Zhou, Licheng Jiao

^*Corresponding author for this work

School of Electronics, Electrical Engineering and Computer Science

Research output: Contribution to journal › Article › peer-review

39 Citations (Scopus)

Abstract

Dimensionality reduction is an essential task in hyperspectral image processing. How to preserve the original intrinsic structure information and enhance the discriminant ability is still a challenge in this area. Recently, with the advantage of preserving global intrinsic structure information, low-rank representation has been applied to dimensionality reduction and achieved promising performance. By exploiting the submanifold information of the original data set, multimanifold learning is effective in enhancing the discriminant ability of the processed data set. In addition, due to the ability of preserving the spatial neighborhood structure information, the tensor analysis has become a popular technique for hyperspectral image processing. Motivated by the above-mentioned analysis, a novel tensor-based low-rank graph with multimanifold regularization (T-LGMR) for dimensionality reduction of hyperspectral images is proposed in this paper. In the T-LGMR, a low-rank constraint is employed to preserve the global data structure while multimanifold information is utilized to enhance the discriminant ability, and tensor representation is used to preserve the spatial neighborhood information. Finally, dimensionality reduction is achieved in the graph embedding framework. Experimental results on three real hyperspectral data sets demonstrate the superiority of the proposed method over several state-of-the-art approaches.

Original language	English
Article number	8383682
Pages (from-to)	4731-4746
Number of pages	16
Journal	IEEE Transactions on Geoscience and Remote Sensing
Volume	56
Issue number	8
DOIs	https://doi.org/10.1109/TGRS.2018.2835514
Publication status	Published - 12 Jun 2018

Bibliographical note

Funding Information:
Manuscript received March 20, 2018; revised May 7, 2018; accepted May 7, 2018. Date of publication June 12, 2018; date of current version July 20, 2018. This work was supported in part by the National Natural Science Foundation of China under Grant 61772400, Grant 61501353, Grant 61772399, Grant 91438201, and Grant 61573267. The work of H. Zhou was supported in part by U.K. EPSRC under Grant EP/N508664/1, Grant EP/R007187/1, and Grant EP/N011074/1, and in part by the Royal Society-Newton Advanced Fellowship under Grant NA160342. (Corresponding author: Xiangrong Zhang.) J. An is with the Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education, Xidian University, Xi’an 710071, China, and also with the School of Information Engineering, Henan Institute of Science and Technology, Xinxiang 453003, China.

Publisher Copyright:
© 1980-2012 IEEE.

Copyright:
Copyright 2018 Elsevier B.V., All rights reserved.

Keywords

Dimensionality reduction
graph embedding
hyperspectral images classification
tensor processing

ASJC Scopus subject areas

Electrical and Electronic Engineering
General Earth and Planetary Sciences

Access to Document

10.1109/TGRS.2018.2835514Licence: Unspecified

Cite this

@article{14b5dc846f6a4596ba73e6c2c56bea8c,

title = "Tensor-based low-rank graph with multimanifold regularization for dimensionality reduction of hyperspectral images",

abstract = "Dimensionality reduction is an essential task in hyperspectral image processing. How to preserve the original intrinsic structure information and enhance the discriminant ability is still a challenge in this area. Recently, with the advantage of preserving global intrinsic structure information, low-rank representation has been applied to dimensionality reduction and achieved promising performance. By exploiting the submanifold information of the original data set, multimanifold learning is effective in enhancing the discriminant ability of the processed data set. In addition, due to the ability of preserving the spatial neighborhood structure information, the tensor analysis has become a popular technique for hyperspectral image processing. Motivated by the above-mentioned analysis, a novel tensor-based low-rank graph with multimanifold regularization (T-LGMR) for dimensionality reduction of hyperspectral images is proposed in this paper. In the T-LGMR, a low-rank constraint is employed to preserve the global data structure while multimanifold information is utilized to enhance the discriminant ability, and tensor representation is used to preserve the spatial neighborhood information. Finally, dimensionality reduction is achieved in the graph embedding framework. Experimental results on three real hyperspectral data sets demonstrate the superiority of the proposed method over several state-of-the-art approaches.",

keywords = "Dimensionality reduction, graph embedding, hyperspectral images classification, tensor processing",

author = "Jinliang An and Xiangrong Zhang and Huiyu Zhou and Licheng Jiao",

note = "Funding Information: Manuscript received March 20, 2018; revised May 7, 2018; accepted May 7, 2018. Date of publication June 12, 2018; date of current version July 20, 2018. This work was supported in part by the National Natural Science Foundation of China under Grant 61772400, Grant 61501353, Grant 61772399, Grant 91438201, and Grant 61573267. The work of H. Zhou was supported in part by U.K. EPSRC under Grant EP/N508664/1, Grant EP/R007187/1, and Grant EP/N011074/1, and in part by the Royal Society-Newton Advanced Fellowship under Grant NA160342. (Corresponding author: Xiangrong Zhang.) J. An is with the Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education, Xidian University, Xi{\textquoteright}an 710071, China, and also with the School of Information Engineering, Henan Institute of Science and Technology, Xinxiang 453003, China. Publisher Copyright: {\textcopyright} 1980-2012 IEEE. Copyright: Copyright 2018 Elsevier B.V., All rights reserved.",

year = "2018",

month = jun,

day = "12",

doi = "10.1109/TGRS.2018.2835514",

language = "English",

volume = "56",

pages = "4731--4746",

journal = "IEEE Transactions on Geoscience and Remote Sensing",

issn = "0196-2892",

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

number = "8",

}

TY - JOUR

T1 - Tensor-based low-rank graph with multimanifold regularization for dimensionality reduction of hyperspectral images

AU - An, Jinliang

AU - Zhang, Xiangrong

AU - Zhou, Huiyu

AU - Jiao, Licheng

N1 - Funding Information: Manuscript received March 20, 2018; revised May 7, 2018; accepted May 7, 2018. Date of publication June 12, 2018; date of current version July 20, 2018. This work was supported in part by the National Natural Science Foundation of China under Grant 61772400, Grant 61501353, Grant 61772399, Grant 91438201, and Grant 61573267. The work of H. Zhou was supported in part by U.K. EPSRC under Grant EP/N508664/1, Grant EP/R007187/1, and Grant EP/N011074/1, and in part by the Royal Society-Newton Advanced Fellowship under Grant NA160342. (Corresponding author: Xiangrong Zhang.) J. An is with the Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education, Xidian University, Xi’an 710071, China, and also with the School of Information Engineering, Henan Institute of Science and Technology, Xinxiang 453003, China. Publisher Copyright: © 1980-2012 IEEE. Copyright: Copyright 2018 Elsevier B.V., All rights reserved.

PY - 2018/6/12

Y1 - 2018/6/12

N2 - Dimensionality reduction is an essential task in hyperspectral image processing. How to preserve the original intrinsic structure information and enhance the discriminant ability is still a challenge in this area. Recently, with the advantage of preserving global intrinsic structure information, low-rank representation has been applied to dimensionality reduction and achieved promising performance. By exploiting the submanifold information of the original data set, multimanifold learning is effective in enhancing the discriminant ability of the processed data set. In addition, due to the ability of preserving the spatial neighborhood structure information, the tensor analysis has become a popular technique for hyperspectral image processing. Motivated by the above-mentioned analysis, a novel tensor-based low-rank graph with multimanifold regularization (T-LGMR) for dimensionality reduction of hyperspectral images is proposed in this paper. In the T-LGMR, a low-rank constraint is employed to preserve the global data structure while multimanifold information is utilized to enhance the discriminant ability, and tensor representation is used to preserve the spatial neighborhood information. Finally, dimensionality reduction is achieved in the graph embedding framework. Experimental results on three real hyperspectral data sets demonstrate the superiority of the proposed method over several state-of-the-art approaches.

AB - Dimensionality reduction is an essential task in hyperspectral image processing. How to preserve the original intrinsic structure information and enhance the discriminant ability is still a challenge in this area. Recently, with the advantage of preserving global intrinsic structure information, low-rank representation has been applied to dimensionality reduction and achieved promising performance. By exploiting the submanifold information of the original data set, multimanifold learning is effective in enhancing the discriminant ability of the processed data set. In addition, due to the ability of preserving the spatial neighborhood structure information, the tensor analysis has become a popular technique for hyperspectral image processing. Motivated by the above-mentioned analysis, a novel tensor-based low-rank graph with multimanifold regularization (T-LGMR) for dimensionality reduction of hyperspectral images is proposed in this paper. In the T-LGMR, a low-rank constraint is employed to preserve the global data structure while multimanifold information is utilized to enhance the discriminant ability, and tensor representation is used to preserve the spatial neighborhood information. Finally, dimensionality reduction is achieved in the graph embedding framework. Experimental results on three real hyperspectral data sets demonstrate the superiority of the proposed method over several state-of-the-art approaches.

KW - Dimensionality reduction

KW - graph embedding

KW - hyperspectral images classification

KW - tensor processing

U2 - 10.1109/TGRS.2018.2835514

DO - 10.1109/TGRS.2018.2835514

M3 - Article

AN - SCOPUS:85048546844

SN - 0196-2892

VL - 56

SP - 4731

EP - 4746

JO - IEEE Transactions on Geoscience and Remote Sensing

JF - IEEE Transactions on Geoscience and Remote Sensing

IS - 8

M1 - 8383682

ER -

Tensor-based low-rank graph with multimanifold regularization for dimensionality reduction of hyperspectral images

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this