Advanced search
1 file | 4.33 MB Add to list

Feature extraction of hyperspectral images with semi-supervised graph learning

Renbo Luo (UGent) , Wenzhi Liao (UGent) , Xin Huang, Youguo Pi and Wilfried Philips (UGent)
Author
Organization
Abstract
We propose a semisupervised graph learning (SEGL) method for feature extraction of hyperspectral remote sensing imagery in this paper. The proposed SEGL method aims to build a semisupervised graph that can maximize the class discrimination and preserve the local neighborhood information by combining labeled and unlabeled samples. In our semisupervised graph, we connect labeled samples according to their label information and unlabeled samples by their nearest neighborhood information. By sorting the mean distance between a unlabeled sample and labeled samples of each class, we connect the unlabeled sample with all labeled samples belonging to its nearest neighborhood class. Moreover, the proposed SEGL better models the actual differences and similarities between samples, by setting different weights to the edges of connected samples. Experimental results on four real hyperspectral images (HSIs) demonstrate the advantages of our method compared to some related feature extraction methods.
Keywords
DIMENSIONALITY REDUCTION, NEAREST-NEIGHBOR, REMOTE-SENSING IMAGES, LINEAR DISCRIMINANT-ANALYSIS, CLASSIFICATION, Classification, feature extraction, graph, hyperspectral images (HSIs), semisupervised

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 4.33 MB

Citation

Please use this url to cite or link to this publication:

MLA
Luo, Renbo et al. “Feature Extraction of Hyperspectral Images with Semi-supervised Graph Learning.” Ed. Jocelyn Chanussot. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING 9.9 (2016): 4389–4399. Print.
APA
Luo, R., Liao, W., Huang, X., Pi, Y., & Philips, W. (2016). Feature extraction of hyperspectral images with semi-supervised graph learning. (J. Chanussot, Ed.)IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 9(9), 4389–4399.
Chicago author-date
Luo, Renbo, Wenzhi Liao, Xin Huang, Youguo Pi, and Wilfried Philips. 2016. “Feature Extraction of Hyperspectral Images with Semi-supervised Graph Learning.” Ed. Jocelyn Chanussot. Ieee Journal of Selected Topics in Applied Earth Observations and Remote Sensing 9 (9): 4389–4399.
Chicago author-date (all authors)
Luo, Renbo, Wenzhi Liao, Xin Huang, Youguo Pi, and Wilfried Philips. 2016. “Feature Extraction of Hyperspectral Images with Semi-supervised Graph Learning.” Ed. Jocelyn Chanussot. Ieee Journal of Selected Topics in Applied Earth Observations and Remote Sensing 9 (9): 4389–4399.
Vancouver
1.
Luo R, Liao W, Huang X, Pi Y, Philips W. Feature extraction of hyperspectral images with semi-supervised graph learning. Chanussot J, editor. IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING. 2016;9(9):4389–99.
IEEE
[1]
R. Luo, W. Liao, X. Huang, Y. Pi, and W. Philips, “Feature extraction of hyperspectral images with semi-supervised graph learning,” IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, vol. 9, no. 9, pp. 4389–4399, 2016.
@article{7064956,
  abstract     = {We propose a semisupervised graph learning (SEGL) method for feature extraction of hyperspectral remote sensing imagery in this paper. The proposed SEGL method aims to build a semisupervised graph that can maximize the class discrimination and preserve the local neighborhood information by combining labeled and unlabeled samples. In our semisupervised graph, we connect labeled samples according to their label information and unlabeled samples by their nearest neighborhood information. By sorting the mean distance between a unlabeled sample and labeled samples of each class, we connect the unlabeled sample with all labeled samples belonging to its nearest neighborhood class. Moreover, the proposed SEGL better models the actual differences and similarities between samples, by setting different weights to the edges of connected samples. Experimental results on four real hyperspectral images (HSIs) demonstrate the advantages of our method compared to some related feature extraction methods.},
  author       = {Luo, Renbo and Liao, Wenzhi and Huang, Xin and Pi, Youguo and Philips, Wilfried},
  editor       = {Chanussot, Jocelyn},
  issn         = {1939-1404},
  journal      = {IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING},
  keywords     = {DIMENSIONALITY REDUCTION,NEAREST-NEIGHBOR,REMOTE-SENSING IMAGES,LINEAR DISCRIMINANT-ANALYSIS,CLASSIFICATION,Classification,feature extraction,graph,hyperspectral images (HSIs),semisupervised},
  language     = {eng},
  number       = {9},
  pages        = {4389--4399},
  title        = {Feature extraction of hyperspectral images with semi-supervised graph learning},
  url          = {http://dx.doi.org/10.1109/JSTARS.2016.2522564},
  volume       = {9},
  year         = {2016},
}

Altmetric
View in Altmetric
Web of Science
Times cited: