Hyperspectral image classification by bootstrap AdaBoost with random decision stumps

Shuji Kawaguchi, Ryuei Nishii

Research output: Contribution to journalArticle

40 Citations (Scopus)

Abstract

We consider a supervised classification of hyperspectral data using AdaBoost with stump functions as base classifiers. We used the bootstrap method without replacement to improve stability and accuracy and to reduce overtraining. We randomly split a data set into two subsets: one for training and the other one for validation. Subsampling and training/validation steps were repeated to derive the final classifier by the majority vote of the classifiers. This method enabled us to estimate variable relevance to the classification. The relevance measure was used to estimate prior probabilities of the variables for random combinations. In numerical experiments with multispectral and hyperspectral data, the proposed method performed extremely well and showed itself to be superior to support vector machines, artificial neural networks, and other well-known classification methods.

Original languageEnglish
Pages (from-to)3845-3851
Number of pages7
JournalIEEE Transactions on Geoscience and Remote Sensing
Volume45
Issue number11
DOIs
Publication statusPublished - Nov 1 2007

Fingerprint

image classification
Adaptive boosting
Image classification
classifiers
Classifiers
education
bootstrapping
estimates
artificial neural network
set theory
Support vector machines
replacement
Neural networks
decision
method
experiment
Experiments

All Science Journal Classification (ASJC) codes

  • Geochemistry and Petrology
  • Geophysics
  • Computers in Earth Sciences
  • Electrical and Electronic Engineering

Cite this

Hyperspectral image classification by bootstrap AdaBoost with random decision stumps. / Kawaguchi, Shuji; Nishii, Ryuei.

In: IEEE Transactions on Geoscience and Remote Sensing, Vol. 45, No. 11, 01.11.2007, p. 3845-3851.

Research output: Contribution to journalArticle

@article{b37f8ccf557b477b903f0aa4d3f4a9a9,
title = "Hyperspectral image classification by bootstrap AdaBoost with random decision stumps",
abstract = "We consider a supervised classification of hyperspectral data using AdaBoost with stump functions as base classifiers. We used the bootstrap method without replacement to improve stability and accuracy and to reduce overtraining. We randomly split a data set into two subsets: one for training and the other one for validation. Subsampling and training/validation steps were repeated to derive the final classifier by the majority vote of the classifiers. This method enabled us to estimate variable relevance to the classification. The relevance measure was used to estimate prior probabilities of the variables for random combinations. In numerical experiments with multispectral and hyperspectral data, the proposed method performed extremely well and showed itself to be superior to support vector machines, artificial neural networks, and other well-known classification methods.",
author = "Shuji Kawaguchi and Ryuei Nishii",
year = "2007",
month = "11",
day = "1",
doi = "10.1109/TGRS.2007.903708",
language = "English",
volume = "45",
pages = "3845--3851",
journal = "IEEE Transactions on Geoscience and Remote Sensing",
issn = "0196-2892",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "11",

}

TY - JOUR

T1 - Hyperspectral image classification by bootstrap AdaBoost with random decision stumps

AU - Kawaguchi, Shuji

AU - Nishii, Ryuei

PY - 2007/11/1

Y1 - 2007/11/1

N2 - We consider a supervised classification of hyperspectral data using AdaBoost with stump functions as base classifiers. We used the bootstrap method without replacement to improve stability and accuracy and to reduce overtraining. We randomly split a data set into two subsets: one for training and the other one for validation. Subsampling and training/validation steps were repeated to derive the final classifier by the majority vote of the classifiers. This method enabled us to estimate variable relevance to the classification. The relevance measure was used to estimate prior probabilities of the variables for random combinations. In numerical experiments with multispectral and hyperspectral data, the proposed method performed extremely well and showed itself to be superior to support vector machines, artificial neural networks, and other well-known classification methods.

AB - We consider a supervised classification of hyperspectral data using AdaBoost with stump functions as base classifiers. We used the bootstrap method without replacement to improve stability and accuracy and to reduce overtraining. We randomly split a data set into two subsets: one for training and the other one for validation. Subsampling and training/validation steps were repeated to derive the final classifier by the majority vote of the classifiers. This method enabled us to estimate variable relevance to the classification. The relevance measure was used to estimate prior probabilities of the variables for random combinations. In numerical experiments with multispectral and hyperspectral data, the proposed method performed extremely well and showed itself to be superior to support vector machines, artificial neural networks, and other well-known classification methods.

UR - http://www.scopus.com/inward/record.url?scp=35648938713&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35648938713&partnerID=8YFLogxK

U2 - 10.1109/TGRS.2007.903708

DO - 10.1109/TGRS.2007.903708

M3 - Article

AN - SCOPUS:35648938713

VL - 45

SP - 3845

EP - 3851

JO - IEEE Transactions on Geoscience and Remote Sensing

JF - IEEE Transactions on Geoscience and Remote Sensing

SN - 0196-2892

IS - 11

ER -