Supervised image classification based on AdaBoost with contextual weak classifiers

Ryuei Nishii, Shinto Eguchi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

AdaBoost, one of machine learning techniques, is employed for supervised classification of land-cover categories of geostatistical data. We introduce contextual classifiers based on neighboring pixels. First, posterior probabilities are calculated at all pixels. Then, averages of the posteriors in various neighborhoods are calculated, and the averages are used as contextual classifiers. Weights for the classifiers can be determined by minimizing the empirical risk with multiclass. Finally, a linear combination of classifier is obtained. The proposed method is applied to artificial multispectral images and shows an excellent performance similar to the MRF-based classifier with much less computation time.

Original languageEnglish
Title of host publication2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004
Pages1467-1470
Number of pages4
Volume2
Publication statusPublished - 2004
Event2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004 - Anchorage, AK, United States
Duration: Sep 20 2004Sep 24 2004

Other

Other2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004
CountryUnited States
CityAnchorage, AK
Period9/20/049/24/04

Fingerprint

Adaptive boosting
Image classification
pixel
Classifiers
multispectral image
image classification
land cover
Pixels
Learning systems
supervised image classification
machine learning
method

All Science Journal Classification (ASJC) codes

  • Geology
  • Software

Cite this

Nishii, R., & Eguchi, S. (2004). Supervised image classification based on AdaBoost with contextual weak classifiers. In 2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004 (Vol. 2, pp. 1467-1470)

Supervised image classification based on AdaBoost with contextual weak classifiers. / Nishii, Ryuei; Eguchi, Shinto.

2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004. Vol. 2 2004. p. 1467-1470.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Nishii, R & Eguchi, S 2004, Supervised image classification based on AdaBoost with contextual weak classifiers. in 2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004. vol. 2, pp. 1467-1470, 2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004, Anchorage, AK, United States, 9/20/04.
Nishii R, Eguchi S. Supervised image classification based on AdaBoost with contextual weak classifiers. In 2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004. Vol. 2. 2004. p. 1467-1470
Nishii, Ryuei ; Eguchi, Shinto. / Supervised image classification based on AdaBoost with contextual weak classifiers. 2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004. Vol. 2 2004. pp. 1467-1470
@inproceedings{7058e4d25215449487f887934ef4a448,
title = "Supervised image classification based on AdaBoost with contextual weak classifiers",
abstract = "AdaBoost, one of machine learning techniques, is employed for supervised classification of land-cover categories of geostatistical data. We introduce contextual classifiers based on neighboring pixels. First, posterior probabilities are calculated at all pixels. Then, averages of the posteriors in various neighborhoods are calculated, and the averages are used as contextual classifiers. Weights for the classifiers can be determined by minimizing the empirical risk with multiclass. Finally, a linear combination of classifier is obtained. The proposed method is applied to artificial multispectral images and shows an excellent performance similar to the MRF-based classifier with much less computation time.",
author = "Ryuei Nishii and Shinto Eguchi",
year = "2004",
language = "English",
volume = "2",
pages = "1467--1470",
booktitle = "2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004",

}

TY - GEN

T1 - Supervised image classification based on AdaBoost with contextual weak classifiers

AU - Nishii, Ryuei

AU - Eguchi, Shinto

PY - 2004

Y1 - 2004

N2 - AdaBoost, one of machine learning techniques, is employed for supervised classification of land-cover categories of geostatistical data. We introduce contextual classifiers based on neighboring pixels. First, posterior probabilities are calculated at all pixels. Then, averages of the posteriors in various neighborhoods are calculated, and the averages are used as contextual classifiers. Weights for the classifiers can be determined by minimizing the empirical risk with multiclass. Finally, a linear combination of classifier is obtained. The proposed method is applied to artificial multispectral images and shows an excellent performance similar to the MRF-based classifier with much less computation time.

AB - AdaBoost, one of machine learning techniques, is employed for supervised classification of land-cover categories of geostatistical data. We introduce contextual classifiers based on neighboring pixels. First, posterior probabilities are calculated at all pixels. Then, averages of the posteriors in various neighborhoods are calculated, and the averages are used as contextual classifiers. Weights for the classifiers can be determined by minimizing the empirical risk with multiclass. Finally, a linear combination of classifier is obtained. The proposed method is applied to artificial multispectral images and shows an excellent performance similar to the MRF-based classifier with much less computation time.

UR - http://www.scopus.com/inward/record.url?scp=15944378910&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=15944378910&partnerID=8YFLogxK

M3 - Conference contribution

VL - 2

SP - 1467

EP - 1470

BT - 2004 IEEE International Geoscience and Remote Sensing Symposium Proceedings: Science for Society: Exploring and Managing a Changing Planet. IGARSS 2004

ER -