Improved feature selection based on a mutual information measure for hyperspectral image classification

Hyperspectral images contain a large amount of information which presents a major challenge for efficient classification. In this paper the information content of each spectral band is analyzed and an improved feature selection technique is proposed for the minimization of dependent information while maximizing the relevancy based on normalized mutual information (NMI). Experimental results are provided for comparisons among some relevant and recentmethods for hyperspectral feature selection in terms of their classification accuracy using real hyperspectral images.

[1]  Kari Torkkola,et al.  Feature Extraction by Non-Parametric Mutual Information Maximization , 2003, J. Mach. Learn. Res..

[2]  E. LeDrew,et al.  Remote sensing of aquatic coastal ecosystem processes , 2006 .

[3]  Keinosuke Fukunaga,et al.  Introduction to Statistical Pattern Recognition , 1972 .

[4]  John A. Richards,et al.  Remote Sensing Digital Image Analysis , 1986 .

[5]  Pao-Ta Yu,et al.  A Nonparametric Feature Extraction and Its Application to Nearest Neighbor Classification for Hyperspectral Image Data , 2010, IEEE Transactions on Geoscience and Remote Sensing.

[6]  Fabio Maselli,et al.  Selection of optimum bands from TM scenes through mutual information analysis , 1993 .

[7]  David A. Landgrebe,et al.  Classification of High Dimensional Data , 1998 .

[8]  Jon Atli Benediktsson,et al.  Kernel Principal Component Analysis for the Classification of Hyperspectral Remote Sensing Data over Urban Areas , 2009, EURASIP J. Adv. Signal Process..

[9]  Fuhui Long,et al.  Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancy , 2003, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[10]  David A. Landgrebe,et al.  Covariance estimation with limited training samples , 1999, IEEE Trans. Geosci. Remote. Sens..

[11]  Roberto Battiti,et al.  Using mutual information for selecting features in supervised neural net learning , 1994, IEEE Trans. Neural Networks.

[12]  Lorenzo Bruzzone,et al.  An extension of the Jeffreys-Matusita distance to multiclass cases for feature selection , 1995, IEEE Trans. Geosci. Remote. Sens..

[13]  G. F. Hughes,et al.  On the mean accuracy of statistical pattern recognizers , 1968, IEEE Trans. Inf. Theory.

[14]  Jacek M. Zurada,et al.  Normalized Mutual Information Feature Selection , 2009, IEEE Transactions on Neural Networks.

[15]  Liu Ying,et al.  Hyperspectral Feature Extraction using Selective PCA based on Genetic Algorithm with Subgroups , 2006, First International Conference on Innovative Computing, Information and Control - Volume I (ICICIC'06).