Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
1 file | 4.09 MB
Add to list
  1. Search results
  2. A robust dynamic classifier selection...

A robust dynamic classifier selection approach for hyperspectral images with imprecise label information

Meizhu Li (UGent) , Shaoguang Huang (UGent) , Jasper De Bock (UGent) , Gert De Cooman (UGent) and Aleksandra Pizurica (UGent)
(2020) SENSORS. 20(18).
Author
Organization
Project
Abstract
Supervised hyperspectral image (HSI) classification relies on accurate label information. However, it is not always possible to collect perfectly accurate labels for training samples. This motivates the development of classifiers that are sufficiently robust to some reasonable amounts of errors in data labels. Despite the growing importance of this aspect, it has not been sufficiently studied in the literature yet. In this paper, we analyze the effect of erroneous sample labels on probability distributions of the principal components of HSIs, and provide in this way a statistical analysis of the resulting uncertainty in classifiers. Building on the theory of imprecise probabilities, we develop a novel robust dynamic classifier selection (R-DCS) model for data classification with erroneous labels. Particularly, spectral and spatial features are extracted from HSIs to construct two individual classifiers for the dynamic selection, respectively. The proposed R-DCS model is based on the robustness of the classifiers’ predictions: the extent to which a classifier can be altered without changing its prediction. We provide three possible selection strategies for the proposed model with different computational complexities and apply them on three benchmark data sets. Experimental results demonstrate that the proposed model outperforms the individual classifiers it selects from and is more robust to errors in labels compared to widely adopted approaches.
Keywords
robust classification, dynamic classifier selection, hyperspectral images, noisy labels, imprecise probabilities, IEEE GRSS DATA, URBAN LAND-USE, DATA FUSION, LIDAR DATA, SEGMENTATION, COMPETENCE, SATELLITE, NOISE, MODEL, RGB

Downloads

  • Download
    sensors-20-05262.pdf
    • full text (Published version)
    • |
    • open access
    • |
    • PDF
    • |
    • 4.09 MB

Citation

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

MLA
Li, Meizhu, et al. “A Robust Dynamic Classifier Selection Approach for Hyperspectral Images with Imprecise Label Information.” SENSORS, vol. 20, no. 18, 2020, doi:10.3390/s20185262.
APA
Li, M., Huang, S., De Bock, J., De Cooman, G., & Pizurica, A. (2020). A robust dynamic classifier selection approach for hyperspectral images with imprecise label information. SENSORS, 20(18). https://doi.org/10.3390/s20185262
Chicago author-date
Li, Meizhu, Shaoguang Huang, Jasper De Bock, Gert De Cooman, and Aleksandra Pizurica. 2020. “A Robust Dynamic Classifier Selection Approach for Hyperspectral Images with Imprecise Label Information.” SENSORS 20 (18). https://doi.org/10.3390/s20185262.
Chicago author-date (all authors)
Li, Meizhu, Shaoguang Huang, Jasper De Bock, Gert De Cooman, and Aleksandra Pizurica. 2020. “A Robust Dynamic Classifier Selection Approach for Hyperspectral Images with Imprecise Label Information.” SENSORS 20 (18). doi:10.3390/s20185262.
Vancouver
1.
Li M, Huang S, De Bock J, De Cooman G, Pizurica A. A robust dynamic classifier selection approach for hyperspectral images with imprecise label information. SENSORS. 2020;20(18).
IEEE
[1]
M. Li, S. Huang, J. De Bock, G. De Cooman, and A. Pizurica, “A robust dynamic classifier selection approach for hyperspectral images with imprecise label information,” SENSORS, vol. 20, no. 18, 2020.
@article{8674714,
  abstract     = {{Supervised hyperspectral image (HSI) classification relies on accurate label information. However, it is not always possible to collect perfectly accurate labels for training samples. This motivates the development of classifiers that are sufficiently robust to some reasonable amounts of errors in data labels. Despite the growing importance of this aspect, it has not been sufficiently studied in the literature yet. In this paper, we analyze the effect of erroneous sample labels on probability distributions of the principal components of HSIs, and provide in this way a statistical analysis of the resulting uncertainty in classifiers. Building on the theory of imprecise probabilities, we develop a novel robust dynamic classifier selection (R-DCS) model for data classification with erroneous labels. Particularly, spectral and spatial features are extracted from HSIs to construct two individual classifiers for the dynamic selection, respectively. The proposed R-DCS model is based on the robustness of the classifiers’ predictions: the extent to which a classifier can be altered without changing its prediction. We provide three possible selection strategies for the proposed model with different computational complexities and apply them on three benchmark data sets. Experimental results demonstrate that the proposed model outperforms the individual classifiers it selects from and is more robust to errors in labels compared to widely adopted approaches.}},
  articleno    = {{5262}},
  author       = {{Li, Meizhu and Huang, Shaoguang and De Bock, Jasper and De Cooman, Gert and Pizurica, Aleksandra}},
  issn         = {{1424-8220}},
  journal      = {{SENSORS}},
  keywords     = {{robust classification,dynamic classifier selection,hyperspectral images,noisy labels,imprecise probabilities,IEEE GRSS DATA,URBAN LAND-USE,DATA FUSION,LIDAR DATA,SEGMENTATION,COMPETENCE,SATELLITE,NOISE,MODEL,RGB}},
  language     = {{eng}},
  number       = {{18}},
  pages        = {{28}},
  title        = {{A robust dynamic classifier selection approach for hyperspectral images with imprecise label information}},
  url          = {{http://doi.org/10.3390/s20185262}},
  volume       = {{20}},
  year         = {{2020}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: