Feature selection and classification of hyperspectral images with support vector machines

Rick Archibald; George Fann

doi:10.1109/LGRS.2007.905116

Feature selection and classification of hyperspectral images with support vector machines

Rick Archibald, George Fann

Data Analysis and Machine Learning

Research output: Contribution to journal › Article › peer-review

242 Scopus citations

Abstract

Hyperspectral images consist of large number of bands which require sophisticated analysis to extract. One approach to reduce computational cost, information representation, and accelerate knowledge discovery is to eliminate bands that do not add value to the classification and analysis method which is being applied. In particular, algorithms that perform band elimination should be designed to take advantage of the structure of the classification method used. This letter introduces an embedded-feature-selection (EFS) algorithm that is tailored to operate with support vector machines (SVMs) to perform band selection and classification simultaneously. We have successfully applied this algorithm to determine a reasonable subset of bands without any user-defined stopping criteria on some sample AVIRIS images; a problem occurs in benchmarking recursive-feature-elimination methods for the SVMs.

Original language	English
Pages (from-to)	674-677
Number of pages	4
Journal	IEEE Geoscience and Remote Sensing Letters
Volume	4
Issue number	4
DOIs	https://doi.org/10.1109/LGRS.2007.905116
State	Published - Oct 2007

Funding

The submitted manuscript has been authored by contractors [UT-Battelle, Manager of Oak Ridge National Laboratory (ORNL)] of the U.S. Government under Contract DE-AC05-00OR22725. Accordingly, the U.S. Government retains a nonexclusive royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so, for U.S. Government purposes. Manuscript received January 17, 2007; revised June 27, 2007. The work of R. Archibald was supported in part by the Householder Fellowship in Scientific Computing sponsored by the U.S. Department of Energy’s Applied Mathematical Sciences Program and in part by the Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC for the U.S. Department of Energy under Contract DE-AC05-00OR22725.

Funders	Funder number
U.S. Government	DE-AC05-00OR22725
U.S. Department of Energy
Oak Ridge National Laboratory
UT-Battelle

Keywords

Feature selection
Hyperspectral images
Support vector machines (SVMs)

Access to Document

10.1109/LGRS.2007.905116

Cite this

@article{2617c88a55164d40aaeb98bed827271d,

title = "Feature selection and classification of hyperspectral images with support vector machines",

abstract = "Hyperspectral images consist of large number of bands which require sophisticated analysis to extract. One approach to reduce computational cost, information representation, and accelerate knowledge discovery is to eliminate bands that do not add value to the classification and analysis method which is being applied. In particular, algorithms that perform band elimination should be designed to take advantage of the structure of the classification method used. This letter introduces an embedded-feature-selection (EFS) algorithm that is tailored to operate with support vector machines (SVMs) to perform band selection and classification simultaneously. We have successfully applied this algorithm to determine a reasonable subset of bands without any user-defined stopping criteria on some sample AVIRIS images; a problem occurs in benchmarking recursive-feature-elimination methods for the SVMs.",

keywords = "Feature selection, Hyperspectral images, Support vector machines (SVMs)",

author = "Rick Archibald and George Fann",

year = "2007",

month = oct,

doi = "10.1109/LGRS.2007.905116",

language = "English",

volume = "4",

pages = "674--677",

journal = "IEEE Geoscience and Remote Sensing Letters",

issn = "1545-598X",

publisher = "Institute of Electrical and Electronics Engineers",

number = "4",

}

TY - JOUR

T1 - Feature selection and classification of hyperspectral images with support vector machines

AU - Archibald, Rick

AU - Fann, George

PY - 2007/10

Y1 - 2007/10

N2 - Hyperspectral images consist of large number of bands which require sophisticated analysis to extract. One approach to reduce computational cost, information representation, and accelerate knowledge discovery is to eliminate bands that do not add value to the classification and analysis method which is being applied. In particular, algorithms that perform band elimination should be designed to take advantage of the structure of the classification method used. This letter introduces an embedded-feature-selection (EFS) algorithm that is tailored to operate with support vector machines (SVMs) to perform band selection and classification simultaneously. We have successfully applied this algorithm to determine a reasonable subset of bands without any user-defined stopping criteria on some sample AVIRIS images; a problem occurs in benchmarking recursive-feature-elimination methods for the SVMs.

AB - Hyperspectral images consist of large number of bands which require sophisticated analysis to extract. One approach to reduce computational cost, information representation, and accelerate knowledge discovery is to eliminate bands that do not add value to the classification and analysis method which is being applied. In particular, algorithms that perform band elimination should be designed to take advantage of the structure of the classification method used. This letter introduces an embedded-feature-selection (EFS) algorithm that is tailored to operate with support vector machines (SVMs) to perform band selection and classification simultaneously. We have successfully applied this algorithm to determine a reasonable subset of bands without any user-defined stopping criteria on some sample AVIRIS images; a problem occurs in benchmarking recursive-feature-elimination methods for the SVMs.

KW - Feature selection

KW - Hyperspectral images

KW - Support vector machines (SVMs)

UR - http://www.scopus.com/inward/record.url?scp=35348920168&partnerID=8YFLogxK

U2 - 10.1109/LGRS.2007.905116

DO - 10.1109/LGRS.2007.905116

M3 - Article

AN - SCOPUS:35348920168

SN - 1545-598X

VL - 4

SP - 674

EP - 677

JO - IEEE Geoscience and Remote Sensing Letters

JF - IEEE Geoscience and Remote Sensing Letters

IS - 4

ER -

Feature selection and classification of hyperspectral images with support vector machines

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this