Advanced search
1 file | 2.76 MB Add to list

Unsupervised spectral sub-feature learning for hyperspectral image classification

Viktor Slavkovikj (UGent) , Steven Verstockt (UGent) , Wesley De Neve (UGent) , Sofie Van Hoecke (UGent) and Rik Van de Walle (UGent)
Author
Organization
Abstract
Spectral pixel classification is one of the principal techniques used in hyperspectral image (HSI) analysis. In this article, we propose an unsupervised feature learning method for classification of hyperspectral images. The proposed method learns a dictionary of sub-feature basis representations from the spectral domain, which allows effective use of the correlated spectral data. The learned dictionary is then used in encoding convolutional samples from the hyperspectral input pixels to an expanded but sparse feature space. Expanded hyperspectral feature representations enable linear separation between object classes present in an image. To evaluate the proposed method, we performed experiments on several commonly used HSI data sets acquired at different locations and by different sensors. Our experimental results show that the proposed method outperforms other pixel-wise classification methods that make use of unsupervised feature extraction approaches. Additionally, even though our approach does not use any prior knowledge, or labelled training data to learn features, it yields either advantageous, or comparable, results in terms of classification accuracy with respect to recent semi-supervised methods.
Keywords
NONLINEAR DIMENSIONALITY REDUCTION, INDEPENDENT COMPONENT ANALYSIS, LINEAR DISCRIMINANT-ANALYSIS, FRAMEWORK, SELECTION, SVM, CONSTRAINT, FACE RECOGNITION, FEATURE-EXTRACTION

Downloads

  • 2016 - Viktor Slavkovikj et al. - Unsupervised spectral sub-feature learning for hyperspectral image classification.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 2.76 MB

Citation

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

MLA
Slavkovikj, Viktor, et al. “Unsupervised Spectral Sub-Feature Learning for Hyperspectral Image Classification.” INTERNATIONAL JOURNAL OF REMOTE SENSING, vol. 37, no. 2, TAYLOR & FRANCIS LTD, 2016, pp. 309–26, doi:10.1080/01431161.2015.1125554.
APA
Slavkovikj, V., Verstockt, S., De Neve, W., Van Hoecke, S., & Van de Walle, R. (2016). Unsupervised spectral sub-feature learning for hyperspectral image classification. INTERNATIONAL JOURNAL OF REMOTE SENSING, 37(2), 309–326. https://doi.org/10.1080/01431161.2015.1125554
Chicago author-date
Slavkovikj, Viktor, Steven Verstockt, Wesley De Neve, Sofie Van Hoecke, and Rik Van de Walle. 2016. “Unsupervised Spectral Sub-Feature Learning for Hyperspectral Image Classification.” INTERNATIONAL JOURNAL OF REMOTE SENSING 37 (2): 309–26. https://doi.org/10.1080/01431161.2015.1125554.
Chicago author-date (all authors)
Slavkovikj, Viktor, Steven Verstockt, Wesley De Neve, Sofie Van Hoecke, and Rik Van de Walle. 2016. “Unsupervised Spectral Sub-Feature Learning for Hyperspectral Image Classification.” INTERNATIONAL JOURNAL OF REMOTE SENSING 37 (2): 309–326. doi:10.1080/01431161.2015.1125554.
Vancouver
1.
Slavkovikj V, Verstockt S, De Neve W, Van Hoecke S, Van de Walle R. Unsupervised spectral sub-feature learning for hyperspectral image classification. INTERNATIONAL JOURNAL OF REMOTE SENSING. 2016;37(2):309–26.
IEEE
[1]
V. Slavkovikj, S. Verstockt, W. De Neve, S. Van Hoecke, and R. Van de Walle, “Unsupervised spectral sub-feature learning for hyperspectral image classification,” INTERNATIONAL JOURNAL OF REMOTE SENSING, vol. 37, no. 2, pp. 309–326, 2016.
@article{7087431,
  abstract     = {{Spectral pixel classification is one of the principal techniques used in hyperspectral image (HSI) analysis. In this article, we propose an unsupervised feature learning method for classification of hyperspectral images. The proposed method learns a dictionary of sub-feature basis representations from the spectral domain, which allows effective use of the correlated spectral data. The learned dictionary is then used in encoding convolutional samples from the hyperspectral input pixels to an expanded but sparse feature space. Expanded hyperspectral feature representations enable linear separation between object classes present in an image. To evaluate the proposed method, we performed experiments on several commonly used HSI data sets acquired at different locations and by different sensors. Our experimental results show that the proposed method outperforms other pixel-wise classification methods that make use of unsupervised feature extraction approaches. Additionally, even though our approach does not use any prior knowledge, or labelled training data to learn features, it yields either advantageous, or comparable, results in terms of classification accuracy with respect to recent semi-supervised methods.}},
  author       = {{Slavkovikj, Viktor and Verstockt, Steven and De Neve, Wesley and Van Hoecke, Sofie and Van de Walle, Rik}},
  issn         = {{0143-1161}},
  journal      = {{INTERNATIONAL JOURNAL OF REMOTE SENSING}},
  keywords     = {{NONLINEAR DIMENSIONALITY REDUCTION,INDEPENDENT COMPONENT ANALYSIS,LINEAR DISCRIMINANT-ANALYSIS,FRAMEWORK,SELECTION,SVM,CONSTRAINT,FACE RECOGNITION,FEATURE-EXTRACTION}},
  language     = {{eng}},
  number       = {{2}},
  pages        = {{309--326}},
  publisher    = {{TAYLOR & FRANCIS LTD}},
  title        = {{Unsupervised spectral sub-feature learning for hyperspectral image classification}},
  url          = {{http://doi.org/10.1080/01431161.2015.1125554}},
  volume       = {{37}},
  year         = {{2016}},
}

Altmetric
View in Altmetric
Web of Science
Times cited: