A Hierarchical Multiclassifier System for Hyperspectral Data Analysis
Many real world classification problems involve high dimensional inputs and a large number of classes. Feature extraction and modular learning approaches can be used to simplify such problems. In this paper, we introduce a hierarchical multiclassifier paradigm in which a C- class problem is recursively decomposed into C- 1 two-class problems. A generalized modular learning framework is used to partition a set of classes into two disjoint groups called meta-classes. The coupled problem of finding a good partition and of searching for a linear feature extractor that best discriminates the resulting two meta-classes are solved simultaneously at each stage of the recursive algorithm. This results in a binary tree whose leaf nodes represent the original C classes. The proposed hierarchical multiclassifier architecture was used to classify 12 types of landcover from 183-dimensional hyperspectral data. The classification accuracy was significantly improved by 4 to 10% relative to other feature extraction and modular learning approaches. Moreover, the class hierarchy that was automatically discovered conformed very well with a human domain expert–s opinion, which demonstrates the potential of such a modular learning approach for discovering domain knowledge automatically from data.
Unable to display preview. Download preview PDF.
- 1.Roderick Murray-Smith and Tor Arne Johansen. Multiple Model Approaches to Modelling and Control. Taylor and Francis, UK, 1997.Google Scholar
- 2.A. Sharkey, editor. Combining Artificial Neural Nets. Springer-Verlag, 1999.Google Scholar
- 3.V. Haertel and D. Landgrebe. On the classification of classes with nearly equal spectral responses in remote sensing hyperspectral image data. IEEE Transactions on Geoscience and Remote Sensing, 37(5–2):2374–2386, September 1999.Google Scholar
- 4.C. Lee and D. A. Landgrebe. Decision boundary feature extraction for neural networks. IEEE Transactions on Neural Networks, 8(1):75–83, January 1997.Google Scholar
- 5.X. Jia and J. A. Richards. Segmented principal components transformation for efficient hyperspectral remote-sensing image display and classification. IEEE Transactions on Geoscience and Remote Sensing, 37(1):538–542, January 1999.Google Scholar
- 6.X. Jia. Classification techniques for hyperspectral remote sensing image data. PhD thesis, Univ. College, ADFA, University of New South Wales, Australia, 1996.Google Scholar
- 7.S. Kumar, J. Ghosh, and M. M. Crawford. Classification of hyperspectral data using best-bases feature extraction algorithms. In Proc. of SPIE: Applications and Science of Computational Intelligence III, Orlando, April 2000.Google Scholar
- 8.N. Saito and Ronald R. Coifman. Local discriminant bases. In Mathematical Imaging: Wavelet Applications in Signal and Image Processing II, Proc. of SPIE, volume 2303, pages 2–14, 1994.Google Scholar
- 9.S. Kumar, M. M. Crawford, and J. Ghosh. A versatile framework for labeling imagery with large number of classes. In Proceedings of the International Joint Conference on Neural Networks, Washington, D.C., 1999.Google Scholar
- 10.M. M. Crawford, S. Kumar, M.R. Ricard, J.C. Gibeaut, and A. Neuenshwander. Fusion of airborne polarimetric and interferometric SAR for classification of coastal environments. IEEE Transactions on Geoscience and Remote Sensing, 37(3):1306–1315, May 1999.Google Scholar
- 12.Yoav Freund and Robert E. Schapire. Experiments with a new boosting algorithm. In Proceedings of the 13th International Conference on Machine Learning, pages 148–156. Morgan Kaufmann, 1996.Google Scholar
- 14.S. Kumar and J. Ghosh. GAMLS: A generalized framework for associative modular learning systems (invited paper). In Proceedings of the Applications and Science of Computational Intelligence II, pages 24–34, Orlando, Florida, 1999.Google Scholar
- 15.R. A. Fisher. The use of multiple measurements in taxonomic problems. Annal of Eugenics, 7:179–188, 1936.Google Scholar