Fusion of SVMs in wavelet domain for hyperspectral data classification

Authors:
Jin Chen;Cheng Wang;Runsheng Wang
Affiliations:
ATR Laboratory, School of Electronic Science and Engineering, National University of Defense Technology, Changsha, China;ATR Laboratory, School of Electronic Science and Engineering, National University of Defense Technology, Changsha, China;ATR Laboratory, School of Electronic Science and Engineering, National University of Defense Technology, Changsha, China
Venue:
ROBIO'09 Proceedings of the 2009 international conference on Robotics and biomimetics
Year:
2009

Citing 5
Cited 0

A Theory for Multiresolution Signal Decomposition: The Wavelet Representation

IEEE Transactions on Pattern Analysis and Machine Intelligence
Probability Estimates for Multi-class Classification by Pairwise Coupling

The Journal of Machine Learning Research
A multiresolution spectral angle-based hyperspectral classification method

International Journal of Remote Sensing
Spectral derivative feature coding for hyperspectral signature analysis

Pattern Recognition
Letters: Adaptive binary tree for fast SVM multiclass classification

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Discrete wavelet transform (DWT) provides a multiresolution view of hyperspectral data. This paper proposes a method to combine the wavelet features at different layers to improve the classification accuracy of hyperspectral data, where both global and local spectral features could be exploited. After feature extraction using DWT, the wavelet feature set of each layer is processed independently by support vector machines (SVMs). Then, the probability outputs of SVMs at each layer are fused to get the final class probability, and the classification result will be the class label with the maximum final class probability. Experimented with the Washington DC Mall hyperspectral data, the results demonstrate that the proposed method can outperform the same classifier with original features, the wavelet features (without fusion), and the wavelet energy features.