Canonical correlation feature selection for sensors with overlapping bands: Theory and application

Biliana Paskaleva, Majeed M. Hayat, Zhipeng Wang, J Scott Tyo, Sanjay Krishna

Research output: Contribution to journalArticle

34 Citations (Scopus)

Abstract

The main focus of this paper is a rigorous development and validation of a novel canonical correlation feature-selection (CCFS) algorithm that is particularly well suited for spectral sensors with overlapping and noisy bands. The proposed approach combines a generalized canonical correlation analysis framework and a minimum mean-square-error criterion for the selection of feature subspaces. The latter induces ranking of the best linear combinations of the noisy overlapping bands and, in doing so, guarantees a minimal generalized distance between the centers of classes and their respective reconstructions in the space spanned by sensor bands. To demonstrate the efficacy and the scope of the proposed approach, two different applications are considered. The first one is separability and classification analysis of rock species using laboratory spectral data and a quantum-dot infrared photodetector (QDIP) sensor. The second application deals with supervised classification and spectral unmixing, and abundance estimation of hyperspectral imagery obtained from the Airborne Hyperspectral Imager sensor. Since QDIP bands exhibit significant spectral overlap, the first study validates the new algorithm in this important application context. The results demonstrate that proper postprocessing can facilitate the emergence of QDIP-based sensors as a promising technology for midwave- and longwave-infrared remote sensing and spectral imaging. In particular, the proposed CCFS algorithm makes it possible to exploit the unique advantage offered by QDIPs with a dot-in-a-well configuration, comprising their bias-dependent spectral response, which is attributable to the quantum Stark effect. The main objective of the second study is to assert that the scope of the new CCFS approach also extends to more traditional spectral sensors.

Original languageEnglish (US)
Article number4637967
Pages (from-to)3346-3358
Number of pages13
JournalIEEE Transactions on Geoscience and Remote Sensing
Volume46
Issue number10
DOIs
StatePublished - Oct 2008

Fingerprint

Feature extraction
sensor
Sensors
Photodetectors
Semiconductor quantum dots
Infrared radiation
abundance estimation
Stark effect
image classification
Image sensors
Mean square error
ranking
Remote sensing
imagery
Rocks
remote sensing
Imaging techniques
rock
analysis

Keywords

  • Canonical correlation (CC) analysis
  • Classification
  • Dot-in-a-well (DWELL)
  • Feature selection
  • Infrared photodetectors
  • Quantum dots
  • Spectral imaging
  • Spectral sensing
  • Subspace projection

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Earth and Planetary Sciences(all)

Cite this

Canonical correlation feature selection for sensors with overlapping bands : Theory and application. / Paskaleva, Biliana; Hayat, Majeed M.; Wang, Zhipeng; Tyo, J Scott; Krishna, Sanjay.

In: IEEE Transactions on Geoscience and Remote Sensing, Vol. 46, No. 10, 4637967, 10.2008, p. 3346-3358.

Research output: Contribution to journalArticle

Paskaleva, Biliana ; Hayat, Majeed M. ; Wang, Zhipeng ; Tyo, J Scott ; Krishna, Sanjay. / Canonical correlation feature selection for sensors with overlapping bands : Theory and application. In: IEEE Transactions on Geoscience and Remote Sensing. 2008 ; Vol. 46, No. 10. pp. 3346-3358.
@article{5dd0ec8aad004b04b61984d30dbf73b5,
title = "Canonical correlation feature selection for sensors with overlapping bands: Theory and application",
abstract = "The main focus of this paper is a rigorous development and validation of a novel canonical correlation feature-selection (CCFS) algorithm that is particularly well suited for spectral sensors with overlapping and noisy bands. The proposed approach combines a generalized canonical correlation analysis framework and a minimum mean-square-error criterion for the selection of feature subspaces. The latter induces ranking of the best linear combinations of the noisy overlapping bands and, in doing so, guarantees a minimal generalized distance between the centers of classes and their respective reconstructions in the space spanned by sensor bands. To demonstrate the efficacy and the scope of the proposed approach, two different applications are considered. The first one is separability and classification analysis of rock species using laboratory spectral data and a quantum-dot infrared photodetector (QDIP) sensor. The second application deals with supervised classification and spectral unmixing, and abundance estimation of hyperspectral imagery obtained from the Airborne Hyperspectral Imager sensor. Since QDIP bands exhibit significant spectral overlap, the first study validates the new algorithm in this important application context. The results demonstrate that proper postprocessing can facilitate the emergence of QDIP-based sensors as a promising technology for midwave- and longwave-infrared remote sensing and spectral imaging. In particular, the proposed CCFS algorithm makes it possible to exploit the unique advantage offered by QDIPs with a dot-in-a-well configuration, comprising their bias-dependent spectral response, which is attributable to the quantum Stark effect. The main objective of the second study is to assert that the scope of the new CCFS approach also extends to more traditional spectral sensors.",
keywords = "Canonical correlation (CC) analysis, Classification, Dot-in-a-well (DWELL), Feature selection, Infrared photodetectors, Quantum dots, Spectral imaging, Spectral sensing, Subspace projection",
author = "Biliana Paskaleva and Hayat, {Majeed M.} and Zhipeng Wang and Tyo, {J Scott} and Sanjay Krishna",
year = "2008",
month = "10",
doi = "10.1109/TGRS.2008.921637",
language = "English (US)",
volume = "46",
pages = "3346--3358",
journal = "IEEE Transactions on Geoscience and Remote Sensing",
issn = "0196-2892",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "10",

}

TY - JOUR

T1 - Canonical correlation feature selection for sensors with overlapping bands

T2 - Theory and application

AU - Paskaleva, Biliana

AU - Hayat, Majeed M.

AU - Wang, Zhipeng

AU - Tyo, J Scott

AU - Krishna, Sanjay

PY - 2008/10

Y1 - 2008/10

N2 - The main focus of this paper is a rigorous development and validation of a novel canonical correlation feature-selection (CCFS) algorithm that is particularly well suited for spectral sensors with overlapping and noisy bands. The proposed approach combines a generalized canonical correlation analysis framework and a minimum mean-square-error criterion for the selection of feature subspaces. The latter induces ranking of the best linear combinations of the noisy overlapping bands and, in doing so, guarantees a minimal generalized distance between the centers of classes and their respective reconstructions in the space spanned by sensor bands. To demonstrate the efficacy and the scope of the proposed approach, two different applications are considered. The first one is separability and classification analysis of rock species using laboratory spectral data and a quantum-dot infrared photodetector (QDIP) sensor. The second application deals with supervised classification and spectral unmixing, and abundance estimation of hyperspectral imagery obtained from the Airborne Hyperspectral Imager sensor. Since QDIP bands exhibit significant spectral overlap, the first study validates the new algorithm in this important application context. The results demonstrate that proper postprocessing can facilitate the emergence of QDIP-based sensors as a promising technology for midwave- and longwave-infrared remote sensing and spectral imaging. In particular, the proposed CCFS algorithm makes it possible to exploit the unique advantage offered by QDIPs with a dot-in-a-well configuration, comprising their bias-dependent spectral response, which is attributable to the quantum Stark effect. The main objective of the second study is to assert that the scope of the new CCFS approach also extends to more traditional spectral sensors.

AB - The main focus of this paper is a rigorous development and validation of a novel canonical correlation feature-selection (CCFS) algorithm that is particularly well suited for spectral sensors with overlapping and noisy bands. The proposed approach combines a generalized canonical correlation analysis framework and a minimum mean-square-error criterion for the selection of feature subspaces. The latter induces ranking of the best linear combinations of the noisy overlapping bands and, in doing so, guarantees a minimal generalized distance between the centers of classes and their respective reconstructions in the space spanned by sensor bands. To demonstrate the efficacy and the scope of the proposed approach, two different applications are considered. The first one is separability and classification analysis of rock species using laboratory spectral data and a quantum-dot infrared photodetector (QDIP) sensor. The second application deals with supervised classification and spectral unmixing, and abundance estimation of hyperspectral imagery obtained from the Airborne Hyperspectral Imager sensor. Since QDIP bands exhibit significant spectral overlap, the first study validates the new algorithm in this important application context. The results demonstrate that proper postprocessing can facilitate the emergence of QDIP-based sensors as a promising technology for midwave- and longwave-infrared remote sensing and spectral imaging. In particular, the proposed CCFS algorithm makes it possible to exploit the unique advantage offered by QDIPs with a dot-in-a-well configuration, comprising their bias-dependent spectral response, which is attributable to the quantum Stark effect. The main objective of the second study is to assert that the scope of the new CCFS approach also extends to more traditional spectral sensors.

KW - Canonical correlation (CC) analysis

KW - Classification

KW - Dot-in-a-well (DWELL)

KW - Feature selection

KW - Infrared photodetectors

KW - Quantum dots

KW - Spectral imaging

KW - Spectral sensing

KW - Subspace projection

UR - http://www.scopus.com/inward/record.url?scp=53849136303&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=53849136303&partnerID=8YFLogxK

U2 - 10.1109/TGRS.2008.921637

DO - 10.1109/TGRS.2008.921637

M3 - Article

AN - SCOPUS:53849136303

VL - 46

SP - 3346

EP - 3358

JO - IEEE Transactions on Geoscience and Remote Sensing

JF - IEEE Transactions on Geoscience and Remote Sensing

SN - 0196-2892

IS - 10

M1 - 4637967

ER -