Abstract

Nonlinear data-driven dimensionality reduction techniques have recently gained popularity due to the emergence of high dimensional data sets. The algorithmic complexity and storage requirements of these techniques, however, can make them prohibitive in resource-limited applications. It is therefore beneficial to reduce the number of exemplar samples required for performing an out-of-sample extension to a test point. In this paper, we propose a novel method for selecting a minimal set of exemplars and performing the out-of-sample extension. In the case of two-class target recognition with Synthetic Aperture radar (SAR) data, we compare the efficacy of the proposed approach with other approaches for selecting a subset of the available training samples. We show that the proposed algorithm outperforms the existing methods by providing low-dimensional embeddings that maintain interclass separability using fewer retained exemplars.

Original languageEnglish (US)
Title of host publicationICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume3
DOIs
StatePublished - 2007
Event2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07 - Honolulu, HI, United States
Duration: Apr 15 2007Apr 20 2007

Other

Other2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07
CountryUnited States
CityHonolulu, HI
Period4/15/074/20/07

Fingerprint

image classification
Image classification
synthetic aperture radar
Synthetic aperture radar
learning
target recognition
radar data
embedding
set theory
resources
education
requirements

Keywords

  • Classification
  • Dimensionality reduction
  • Out-of-sample extension
  • Reduced complexity isomap
  • SAR

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing
  • Acoustics and Ultrasonics

Cite this

Berisha, V., Shah, N., Waagen, D., Schmitt, H., Bellofiore, S., Spanias, A., & Cochran, D. (2007). Sparse manifold learning with applications to SAR image classification. In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings (Vol. 3). [4217903] https://doi.org/10.1109/ICASSP.2007.366873

Sparse manifold learning with applications to SAR image classification. / Berisha, Visar; Shah, N.; Waagen, D.; Schmitt, H.; Bellofiore, S.; Spanias, Andreas; Cochran, Douglas.

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. Vol. 3 2007. 4217903.

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

Berisha, V, Shah, N, Waagen, D, Schmitt, H, Bellofiore, S, Spanias, A & Cochran, D 2007, Sparse manifold learning with applications to SAR image classification. in ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. vol. 3, 4217903, 2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07, Honolulu, HI, United States, 4/15/07. https://doi.org/10.1109/ICASSP.2007.366873
Berisha V, Shah N, Waagen D, Schmitt H, Bellofiore S, Spanias A et al. Sparse manifold learning with applications to SAR image classification. In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. Vol. 3. 2007. 4217903 https://doi.org/10.1109/ICASSP.2007.366873
Berisha, Visar ; Shah, N. ; Waagen, D. ; Schmitt, H. ; Bellofiore, S. ; Spanias, Andreas ; Cochran, Douglas. / Sparse manifold learning with applications to SAR image classification. ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. Vol. 3 2007.
@inproceedings{a9740765bad543e4bd28a370ccbfc9cc,
title = "Sparse manifold learning with applications to SAR image classification",
abstract = "Nonlinear data-driven dimensionality reduction techniques have recently gained popularity due to the emergence of high dimensional data sets. The algorithmic complexity and storage requirements of these techniques, however, can make them prohibitive in resource-limited applications. It is therefore beneficial to reduce the number of exemplar samples required for performing an out-of-sample extension to a test point. In this paper, we propose a novel method for selecting a minimal set of exemplars and performing the out-of-sample extension. In the case of two-class target recognition with Synthetic Aperture radar (SAR) data, we compare the efficacy of the proposed approach with other approaches for selecting a subset of the available training samples. We show that the proposed algorithm outperforms the existing methods by providing low-dimensional embeddings that maintain interclass separability using fewer retained exemplars.",
keywords = "Classification, Dimensionality reduction, Out-of-sample extension, Reduced complexity isomap, SAR",
author = "Visar Berisha and N. Shah and D. Waagen and H. Schmitt and S. Bellofiore and Andreas Spanias and Douglas Cochran",
year = "2007",
doi = "10.1109/ICASSP.2007.366873",
language = "English (US)",
isbn = "1424407281",
volume = "3",
booktitle = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

}

TY - GEN

T1 - Sparse manifold learning with applications to SAR image classification

AU - Berisha, Visar

AU - Shah, N.

AU - Waagen, D.

AU - Schmitt, H.

AU - Bellofiore, S.

AU - Spanias, Andreas

AU - Cochran, Douglas

PY - 2007

Y1 - 2007

N2 - Nonlinear data-driven dimensionality reduction techniques have recently gained popularity due to the emergence of high dimensional data sets. The algorithmic complexity and storage requirements of these techniques, however, can make them prohibitive in resource-limited applications. It is therefore beneficial to reduce the number of exemplar samples required for performing an out-of-sample extension to a test point. In this paper, we propose a novel method for selecting a minimal set of exemplars and performing the out-of-sample extension. In the case of two-class target recognition with Synthetic Aperture radar (SAR) data, we compare the efficacy of the proposed approach with other approaches for selecting a subset of the available training samples. We show that the proposed algorithm outperforms the existing methods by providing low-dimensional embeddings that maintain interclass separability using fewer retained exemplars.

AB - Nonlinear data-driven dimensionality reduction techniques have recently gained popularity due to the emergence of high dimensional data sets. The algorithmic complexity and storage requirements of these techniques, however, can make them prohibitive in resource-limited applications. It is therefore beneficial to reduce the number of exemplar samples required for performing an out-of-sample extension to a test point. In this paper, we propose a novel method for selecting a minimal set of exemplars and performing the out-of-sample extension. In the case of two-class target recognition with Synthetic Aperture radar (SAR) data, we compare the efficacy of the proposed approach with other approaches for selecting a subset of the available training samples. We show that the proposed algorithm outperforms the existing methods by providing low-dimensional embeddings that maintain interclass separability using fewer retained exemplars.

KW - Classification

KW - Dimensionality reduction

KW - Out-of-sample extension

KW - Reduced complexity isomap

KW - SAR

UR - http://www.scopus.com/inward/record.url?scp=34547514160&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34547514160&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2007.366873

DO - 10.1109/ICASSP.2007.366873

M3 - Conference contribution

AN - SCOPUS:34547514160

SN - 1424407281

SN - 9781424407286

VL - 3

BT - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

ER -