Adaptive Feature Spaces for Land Cover Classification with Limited Ground Truth Data
Classification of hyperspectral data is challenging because of high dimensionality (O(100)) inputs, several possible output classes with uneven priors, and scarcity of labeled information. In an earlier work, a multiclassifier system arranged as a binary hierarchy was developed to group classes for easier, progressive discrimination . This paper substantially expands the scope of such a system by integrating a feature reduction scheme that adaptively adjusts to the amount of labeled data available, while exploiting the highly correlated nature of certain adjacent hyperspectral bands. The resulting best-basis binary hierarchical classifier (BB-BHC) family is thus able to address the “small sample size” problem, as evidenced by our experimental results.
Unable to display preview. Download preview PDF.
- 2.D. Landgrebe, “Information extraction principles and methods for multispectral and hyperspectral image data,” Information Processing for Remote Sensing, ed. Chen, C.H., World Scientific Pub. Co, NJ, 1999.Google Scholar
- 4.S. Kumar, J. Ghosh and M. M. Crawford, “Hierarchical fusion of multiple classifiers for hyperspectral data analysis,” Pattern Analysis and Applications, Special Issue on Classifier Fusion (to appear).Google Scholar
- 6.P.A. Devijver and J. Kittler (editors), Pattern Recognition Theory and Application. Springer-Verlag, 1987.Google Scholar
- 8.S.J. Raudys and A.K. Jain, “Small sample size effects in statistical pattern recognition: recommendations for practitioners”, IEEE Trans on PAMI, 13(3): 252–64, 1991.Google Scholar
- 12.Marina Skurichina, “Stabilizing weak classifiers,” Thesis, Vilnius State University, 2001.Google Scholar
- 14.A. McCallum, R. Rosenfeld, T. Mitchell, and A.Y. Ng, “Improving text classification by shrinkage in a hierarchy of classes,” Proc. 15th International Conf. on Machine, Madison, WI, Morgan Kaufmann, San Mateo, CA, 359–67 1998.Google Scholar
- 15.K. Fukunaga, Introduction to Statistical Pattern Recognition, 2nd Ed, Boston, 1990.Google Scholar
- 18.A. Blum and T. Mitchell, “Combining labeled and unlabeled data with co-training,” Proc. 11 th Annual Conf. Computational Learning Theory, 92–100, 1998.Google Scholar
- 19.Webpage. Jet Propulsion Lab, California Institute of Technology, http://makalu.jpl.nasa.gov/.
- 20.B. Jeon and D. Landgrebe, “Partially supervised classification using weighted unsupervised clustering,” IEEE Trans. Geosci.Rem. Sens., 37(2): 1073–9, March 1999.Google Scholar
- 21.T.M. Mitchell, “The role of unlabeled data in supervised learning,” Proc. Sixth Intl. Colloquium on Cognitive Science, 8 pgs, 1999.Google Scholar
- 22.V.R. de Sa, “Learning classification with unlabeled data,” Advances in Neural Information Processing Systems 6, 1994.Google Scholar
- 24.T. Cocks, R. Jenssen, A. Stewart, I. Wilson, and T. Shields, “The HyMap airborne hyperspectral sensor: the system, calibration and performance”, Proc. 1st EARSeL Workshop on Imaging Spectroscopy (M. Schaepman, D. Schläpfer, and K.I. Itten, Eds.), Zurich, EARSeL, Paris, 37–42, 6–8 October, 1998.Google Scholar
- 25.X. Jia, Classification Techniques for Hyperspectral Remote Sensing Image Data. PhD Thesis, Univ. College, ADFA, University of New South Wales, Australia, 1996.Google Scholar
- 27.S. Kumar, J. Ghosh, and M. M. Crawford, “A hierarchical multiclassifier system for hyperspectral data analysis”, 1st Intl.Workshop on Multiple Classifier Systems, Sardinia, Italy, 270–9, June 2000.Google Scholar
- 29.K. Turner and J. Ghosh, “Error correlation and error reduction in ensemble classifiers” Connection Science, Special Issue on Combining, 8(3/4), 385–404, 1996.Google Scholar