Research on supervised LPP feature extraction for hyperspectral image

Renbo Luo, Youguo Pi, Wenzhi Liao

Research output: Contribution to journalArticle

Abstract

For the classification among different land-cover types in a hyperspectral image, particularly in the small-sample-size problem, a feature extraction method is an approach for reducing the dimensionality and increasing the classification accuracy. A supervised principal locality preserving projection (SPLPP ) feature extraction algorithms, which uses the label information of training sample in locality preserving projection (LPP), was proposed in this paper. Three main steps are involved in the proposed SLPP: firstly uses PCA to remove redundant information, and then combines the label information in LPP, finally, SPLPP projects high-dimensional hyperspectral image into a low-dimensional space. Last but not least, SPLPP uses the extracted features as inputs of classifiers (e.g. support vector machine (SVM ) and K-nearest neighbors (KNN )) to do classification. Experimental results show that the proposed SPLPP has better local information retention ability and class discrimination ability compared with PCA, LPP, LDA
LanguageEnglish
Pages46-52
Number of pages7
JournalREMOTE SENSING TECHNOLOGY AND APPLICATION
Volume27
Issue number6
Publication statusPublished - 2012

Fingerprint

pattern recognition
preserving
projection
extraction method
land cover
classifiers
discrimination
education

Keywords

  • hyperspectral remote sensing image
  • supervised principal locality preserving projection
  • classification
  • feature extraction

Cite this

Luo, Renbo ; Pi, Youguo ; Liao, Wenzhi. / Research on supervised LPP feature extraction for hyperspectral image. 2012 ; Vol. 27, No. 6. pp. 46-52.
@article{98c58b3265e947538e0c316b43828044,
title = "Research on supervised LPP feature extraction for hyperspectral image",
abstract = "For the classification among different land-cover types in a hyperspectral image, particularly in the small-sample-size problem, a feature extraction method is an approach for reducing the dimensionality and increasing the classification accuracy. A supervised principal locality preserving projection (SPLPP ) feature extraction algorithms, which uses the label information of training sample in locality preserving projection (LPP), was proposed in this paper. Three main steps are involved in the proposed SLPP: firstly uses PCA to remove redundant information, and then combines the label information in LPP, finally, SPLPP projects high-dimensional hyperspectral image into a low-dimensional space. Last but not least, SPLPP uses the extracted features as inputs of classifiers (e.g. support vector machine (SVM ) and K-nearest neighbors (KNN )) to do classification. Experimental results show that the proposed SPLPP has better local information retention ability and class discrimination ability compared with PCA, LPP, LDA",
keywords = "hyperspectral remote sensing image, supervised principal locality preserving projection, classification, feature extraction",
author = "Renbo Luo and Youguo Pi and Wenzhi Liao",
year = "2012",
language = "English",
volume = "27",
pages = "46--52",
number = "6",

}

Research on supervised LPP feature extraction for hyperspectral image. / Luo, Renbo; Pi, Youguo; Liao, Wenzhi.

Vol. 27, No. 6, 2012, p. 46-52.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Research on supervised LPP feature extraction for hyperspectral image

AU - Luo, Renbo

AU - Pi, Youguo

AU - Liao, Wenzhi

PY - 2012

Y1 - 2012

N2 - For the classification among different land-cover types in a hyperspectral image, particularly in the small-sample-size problem, a feature extraction method is an approach for reducing the dimensionality and increasing the classification accuracy. A supervised principal locality preserving projection (SPLPP ) feature extraction algorithms, which uses the label information of training sample in locality preserving projection (LPP), was proposed in this paper. Three main steps are involved in the proposed SLPP: firstly uses PCA to remove redundant information, and then combines the label information in LPP, finally, SPLPP projects high-dimensional hyperspectral image into a low-dimensional space. Last but not least, SPLPP uses the extracted features as inputs of classifiers (e.g. support vector machine (SVM ) and K-nearest neighbors (KNN )) to do classification. Experimental results show that the proposed SPLPP has better local information retention ability and class discrimination ability compared with PCA, LPP, LDA

AB - For the classification among different land-cover types in a hyperspectral image, particularly in the small-sample-size problem, a feature extraction method is an approach for reducing the dimensionality and increasing the classification accuracy. A supervised principal locality preserving projection (SPLPP ) feature extraction algorithms, which uses the label information of training sample in locality preserving projection (LPP), was proposed in this paper. Three main steps are involved in the proposed SLPP: firstly uses PCA to remove redundant information, and then combines the label information in LPP, finally, SPLPP projects high-dimensional hyperspectral image into a low-dimensional space. Last but not least, SPLPP uses the extracted features as inputs of classifiers (e.g. support vector machine (SVM ) and K-nearest neighbors (KNN )) to do classification. Experimental results show that the proposed SPLPP has better local information retention ability and class discrimination ability compared with PCA, LPP, LDA

KW - hyperspectral remote sensing image

KW - supervised principal locality preserving projection

KW - classification

KW - feature extraction

UR - http://hdl.handle.net/1854/LU-5840655

M3 - Article

VL - 27

SP - 46

EP - 52

IS - 6

ER -