Supervised Hyperspectral Image Classification with Rejection

Filipe Condessa, Jose Bioucas-Dias, Jelena Kovacevic

Research output: Contribution to journalArticle

Abstract

Hyperspectral image classification is a challenging problem as obtaining complete and representative training sets is costly, pixels can belong to unknown classes, and it is generally an ill-posed problem. The need to achieve high classification accuracy may surpass the need to classify the entire image. To account for this scenario, we use classification with rejection by providing the classifier with an option not to classify a pixel and consequently reject it. We present and analyze two approaches for supervised hyperspectral image classification that combine the use of contextual priors with classification with rejection: 1) by jointly computing context and rejection and 2) by sequentially computing context and rejection. In the joint approach, rejection is introduced as an extra class that models the probability of classifier failure. In the sequential approach, rejection results from the hidden field associated with a marginal maximum a posteriori classification of the image. We validate both approaches on real hyperspectral data.

Original languageEnglish (US)
Article number7393457
Pages (from-to)2321-2332
Number of pages12
JournalIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Volume9
Issue number6
DOIs
StatePublished - Jun 1 2016

Fingerprint

Image classification
pixel
Classifiers
Pixels
image classification
supervised image classification

Keywords

  • Classification with context
  • classification with rejection
  • hyperspectral image classification

ASJC Scopus subject areas

  • Computers in Earth Sciences
  • Atmospheric Science

Cite this

Supervised Hyperspectral Image Classification with Rejection. / Condessa, Filipe; Bioucas-Dias, Jose; Kovacevic, Jelena.

In: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 9, No. 6, 7393457, 01.06.2016, p. 2321-2332.

Research output: Contribution to journalArticle

@article{45092ff99978483b84340e618ab0706c,
title = "Supervised Hyperspectral Image Classification with Rejection",
abstract = "Hyperspectral image classification is a challenging problem as obtaining complete and representative training sets is costly, pixels can belong to unknown classes, and it is generally an ill-posed problem. The need to achieve high classification accuracy may surpass the need to classify the entire image. To account for this scenario, we use classification with rejection by providing the classifier with an option not to classify a pixel and consequently reject it. We present and analyze two approaches for supervised hyperspectral image classification that combine the use of contextual priors with classification with rejection: 1) by jointly computing context and rejection and 2) by sequentially computing context and rejection. In the joint approach, rejection is introduced as an extra class that models the probability of classifier failure. In the sequential approach, rejection results from the hidden field associated with a marginal maximum a posteriori classification of the image. We validate both approaches on real hyperspectral data.",
keywords = "Classification with context, classification with rejection, hyperspectral image classification",
author = "Filipe Condessa and Jose Bioucas-Dias and Jelena Kovacevic",
year = "2016",
month = "6",
day = "1",
doi = "10.1109/JSTARS.2015.2510032",
language = "English (US)",
volume = "9",
pages = "2321--2332",
journal = "IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing",
issn = "1939-1404",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

TY - JOUR

T1 - Supervised Hyperspectral Image Classification with Rejection

AU - Condessa, Filipe

AU - Bioucas-Dias, Jose

AU - Kovacevic, Jelena

PY - 2016/6/1

Y1 - 2016/6/1

N2 - Hyperspectral image classification is a challenging problem as obtaining complete and representative training sets is costly, pixels can belong to unknown classes, and it is generally an ill-posed problem. The need to achieve high classification accuracy may surpass the need to classify the entire image. To account for this scenario, we use classification with rejection by providing the classifier with an option not to classify a pixel and consequently reject it. We present and analyze two approaches for supervised hyperspectral image classification that combine the use of contextual priors with classification with rejection: 1) by jointly computing context and rejection and 2) by sequentially computing context and rejection. In the joint approach, rejection is introduced as an extra class that models the probability of classifier failure. In the sequential approach, rejection results from the hidden field associated with a marginal maximum a posteriori classification of the image. We validate both approaches on real hyperspectral data.

AB - Hyperspectral image classification is a challenging problem as obtaining complete and representative training sets is costly, pixels can belong to unknown classes, and it is generally an ill-posed problem. The need to achieve high classification accuracy may surpass the need to classify the entire image. To account for this scenario, we use classification with rejection by providing the classifier with an option not to classify a pixel and consequently reject it. We present and analyze two approaches for supervised hyperspectral image classification that combine the use of contextual priors with classification with rejection: 1) by jointly computing context and rejection and 2) by sequentially computing context and rejection. In the joint approach, rejection is introduced as an extra class that models the probability of classifier failure. In the sequential approach, rejection results from the hidden field associated with a marginal maximum a posteriori classification of the image. We validate both approaches on real hyperspectral data.

KW - Classification with context

KW - classification with rejection

KW - hyperspectral image classification

UR - http://www.scopus.com/inward/record.url?scp=84961367501&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961367501&partnerID=8YFLogxK

U2 - 10.1109/JSTARS.2015.2510032

DO - 10.1109/JSTARS.2015.2510032

M3 - Article

VL - 9

SP - 2321

EP - 2332

JO - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

JF - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

SN - 1939-1404

IS - 6

M1 - 7393457

ER -