Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images

Muhammad Ahmad, Asad Khan, Adil Mehmood Khan, Manuel Mazzara, Salvatore Distefano, Ahmed Sohaib, O Nibouche

Research output: Contribution to journalArticle

Abstract

Acquisition of labeled data for supervised Hyperspectral Image (HSI) classification is expensive in terms of both time and costs. Moreover, manual selection and labeling are often subjective and tend to induce redundancy into the classifier. Active learning (AL) can be a suitable approach for HSI classification as it integrates data acquisition to the classifier design by ranking the unlabeled data to provide advice for the next query that has the highest training utility. However, multiclass AL techniques tend to include redundant samples into the classifier to some extent. This paper addresses such a problem by introducing an AL pipeline which preserves the most representative and spatially heterogeneous samples. The adopted strategy for sample selection utilizes fuzziness to assess the mapping between actual output and the approximated a-posteriori probabilities, computed by a marginal probability distribution based on discriminative random fields. The samples selected in each iteration are then provided to the spectral angle mapper-based objective function to reduce the inter-class redundancy. Experiments on five HSI benchmark datasets confirmed that the proposed Fuzziness and Spectral Angle Mapper (FSAM)-AL pipeline presents competitive results compared to the state-of-the-art sample selection techniques, leading to lower computational requirements.
LanguageEnglish
Article number1136
Pages1-19
Number of pages19
JournalMDPI Remote Sensing
Volume11
Issue number9
DOIs
Publication statusPublished - 13 May 2019

Fingerprint

learning
image classification
data acquisition
ranking
cost
experiment

Keywords

  • Active learning
  • Fuzziness
  • Hyperspectral imaging
  • Soft threshold
  • Spectral angle mapper

Cite this

Ahmad, M., Khan, A., Khan, A. M., Mazzara, M., Distefano, S., Sohaib, A., & Nibouche, O. (2019). Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images. 11(9), 1-19. [1136]. https://doi.org/10.3390/rs11091136
Ahmad, Muhammad ; Khan, Asad ; Khan, Adil Mehmood ; Mazzara, Manuel ; Distefano, Salvatore ; Sohaib, Ahmed ; Nibouche, O. / Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images. 2019 ; Vol. 11, No. 9. pp. 1-19.
@article{c8da0afe205e417d960b222b0306df2f,
title = "Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images",
abstract = "Acquisition of labeled data for supervised Hyperspectral Image (HSI) classification is expensive in terms of both time and costs. Moreover, manual selection and labeling are often subjective and tend to induce redundancy into the classifier. Active learning (AL) can be a suitable approach for HSI classification as it integrates data acquisition to the classifier design by ranking the unlabeled data to provide advice for the next query that has the highest training utility. However, multiclass AL techniques tend to include redundant samples into the classifier to some extent. This paper addresses such a problem by introducing an AL pipeline which preserves the most representative and spatially heterogeneous samples. The adopted strategy for sample selection utilizes fuzziness to assess the mapping between actual output and the approximated a-posteriori probabilities, computed by a marginal probability distribution based on discriminative random fields. The samples selected in each iteration are then provided to the spectral angle mapper-based objective function to reduce the inter-class redundancy. Experiments on five HSI benchmark datasets confirmed that the proposed Fuzziness and Spectral Angle Mapper (FSAM)-AL pipeline presents competitive results compared to the state-of-the-art sample selection techniques, leading to lower computational requirements.",
keywords = "Active learning, Fuzziness, Hyperspectral imaging, Soft threshold, Spectral angle mapper",
author = "Muhammad Ahmad and Asad Khan and Khan, {Adil Mehmood} and Manuel Mazzara and Salvatore Distefano and Ahmed Sohaib and O Nibouche",
year = "2019",
month = "5",
day = "13",
doi = "10.3390/rs11091136",
language = "English",
volume = "11",
pages = "1--19",
number = "9",

}

Ahmad, M, Khan, A, Khan, AM, Mazzara, M, Distefano, S, Sohaib, A & Nibouche, O 2019, 'Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images', vol. 11, no. 9, 1136, pp. 1-19. https://doi.org/10.3390/rs11091136

Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images. / Ahmad, Muhammad; Khan, Asad; Khan, Adil Mehmood; Mazzara, Manuel; Distefano, Salvatore; Sohaib, Ahmed; Nibouche, O.

Vol. 11, No. 9, 1136, 13.05.2019, p. 1-19.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images

AU - Ahmad, Muhammad

AU - Khan, Asad

AU - Khan, Adil Mehmood

AU - Mazzara, Manuel

AU - Distefano, Salvatore

AU - Sohaib, Ahmed

AU - Nibouche, O

PY - 2019/5/13

Y1 - 2019/5/13

N2 - Acquisition of labeled data for supervised Hyperspectral Image (HSI) classification is expensive in terms of both time and costs. Moreover, manual selection and labeling are often subjective and tend to induce redundancy into the classifier. Active learning (AL) can be a suitable approach for HSI classification as it integrates data acquisition to the classifier design by ranking the unlabeled data to provide advice for the next query that has the highest training utility. However, multiclass AL techniques tend to include redundant samples into the classifier to some extent. This paper addresses such a problem by introducing an AL pipeline which preserves the most representative and spatially heterogeneous samples. The adopted strategy for sample selection utilizes fuzziness to assess the mapping between actual output and the approximated a-posteriori probabilities, computed by a marginal probability distribution based on discriminative random fields. The samples selected in each iteration are then provided to the spectral angle mapper-based objective function to reduce the inter-class redundancy. Experiments on five HSI benchmark datasets confirmed that the proposed Fuzziness and Spectral Angle Mapper (FSAM)-AL pipeline presents competitive results compared to the state-of-the-art sample selection techniques, leading to lower computational requirements.

AB - Acquisition of labeled data for supervised Hyperspectral Image (HSI) classification is expensive in terms of both time and costs. Moreover, manual selection and labeling are often subjective and tend to induce redundancy into the classifier. Active learning (AL) can be a suitable approach for HSI classification as it integrates data acquisition to the classifier design by ranking the unlabeled data to provide advice for the next query that has the highest training utility. However, multiclass AL techniques tend to include redundant samples into the classifier to some extent. This paper addresses such a problem by introducing an AL pipeline which preserves the most representative and spatially heterogeneous samples. The adopted strategy for sample selection utilizes fuzziness to assess the mapping between actual output and the approximated a-posteriori probabilities, computed by a marginal probability distribution based on discriminative random fields. The samples selected in each iteration are then provided to the spectral angle mapper-based objective function to reduce the inter-class redundancy. Experiments on five HSI benchmark datasets confirmed that the proposed Fuzziness and Spectral Angle Mapper (FSAM)-AL pipeline presents competitive results compared to the state-of-the-art sample selection techniques, leading to lower computational requirements.

KW - Active learning

KW - Fuzziness

KW - Hyperspectral imaging

KW - Soft threshold

KW - Spectral angle mapper

UR - http://www.scopus.com/inward/record.url?scp=85065706058&partnerID=8YFLogxK

U2 - 10.3390/rs11091136

DO - 10.3390/rs11091136

M3 - Article

VL - 11

SP - 1

EP - 19

IS - 9

M1 - 1136

ER -

Ahmad M, Khan A, Khan AM, Mazzara M, Distefano S, Sohaib A et al. Spatial Prior Fuzziness Pool-based Interactive Classification of Hyperspectral Images. 2019 May 13;11(9):1-19. 1136. https://doi.org/10.3390/rs11091136