A modular eigen subspace scheme for high-dimensional data classification with NASA MODIS/ASTER (MASTER) airborne simulator data sets of Pacrim II project

被引:4
作者
Chang, YL [1 ]
Han, CC [1 ]
Fan, KC [1 ]
Chen, KS [1 ]
Chang, JH [1 ]
机构
[1] St Johns & St Marys Inst Tech, Dept Comp Sci & Info Eng, Taipei, Taiwan
来源
IMAGING SPECTROMETRY VIII | 2002年 / 4816卷
关键词
high-dimensional supervised classification; positive Boolean function; stack filter; greedy modular subspaces; principal components transformation;
D O I
10.1117/12.451532
中图分类号
O43 [光学];
学科分类号
070207 ; 0803 ;
摘要
In this paper, a novel filter-based greedy modular subspace (GMS) technique is proposed to improve the accuracy of high-dimensional remote sensing image supervisor classification. The approach initially divides the whole set of high-dimensional features into several arbitrary number of highly correlated subgroup by performing a greedy correlation matrix reordering transformation for each class. These GMS can be regarded as a unique feature for each distinguishable class in high-dimensional data sets. The similarity measures are next calculated by projecting the samples into different modular feature subspaces. Finally, a supervised multi-class classifier which is implemented based on positive Boolean function (PBF) schemes is adopted to build, a non-linear optimal classifier. A PBF is exactly one sum-of-product form without any negative components. The PBF possesses the well-known threshold decomposition and stacking properties. The classification errors can be calculated from the summation of the absolute errors incurred at each level. The optimal PBF are found and designed as a classifier by minimize the classification error rate among the training samples. Experimental results demonstrate that the proposed GMS feature extraction method suits the PBF classifier best as a classification preprocess. It significantly improves the precision of image classification compared with conventional feature extraction schemes. Moreover, a practicable and convenient "vague" boundary sampling property of PBF is introduced to visually select training samples from high-dimensional data sets more efficiently.
引用
收藏
页码:426 / 436
页数:11
相关论文
共 16 条
[1]  
[Anonymous], IEEE INT C COMP VIS
[2]   STACK FILTERS AND THE MEAN ABSOLUTE ERROR CRITERION [J].
COYLE, EJ ;
LIN, JH .
IEEE TRANSACTIONS ON ACOUSTICS SPEECH AND SIGNAL PROCESSING, 1988, 36 (08) :1244-1254
[3]   MEDIAN FILTERING BY THRESHOLD DECOMPOSITION [J].
FITCH, JP ;
COYLE, EJ ;
GALLAGHER, NC .
IEEE TRANSACTIONS ON ACOUSTICS SPEECH AND SIGNAL PROCESSING, 1984, 32 (06) :1183-1188
[4]   Finding of optimal stack filter by graphic searching methods [J].
Han, CC ;
Fan, KC .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 1997, 45 (07) :1857-1862
[5]  
HAN CC, 2001, IEEE INT CARN C SEC, P278
[6]  
HAN CC, 2002, IN PRESS INT C PATT
[7]   The MODIS/ASTER airborne simulator (MASTER) - a new instrument for earth science studies [J].
Hook, SJ ;
Myers, JEJ ;
Thome, KJ ;
Fitzgerald, M ;
Kahle, AB .
REMOTE SENSING OF ENVIRONMENT, 2001, 76 (01) :93-102
[8]   Segmented principal components transformation for efficient hyperspectral remote-sensing image display and classification [J].
Jia, XP ;
Richards, JA .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 1999, 37 (01) :538-542
[9]   Hyperspectral data analysis and supervised feature reduction via projection pursuit [J].
Jimenez, LO ;
Landgrebe, DA .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 1999, 37 (06) :2653-2667
[10]   Minimum classification error rate methods for speech recognition [J].
Juang, BH ;
Chou, W ;
Lee, CH .
IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, 1997, 5 (03) :257-265