Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors

G. Székely, Mary Lou Padgett, Gerry Dozier

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

The focus of this paper is on real-world applications of neural networks and/or evolutionary computation, and on recent theoretical developments in parameter selection for pulse coupled neural networks (PCNN's) applicable to automatic target recognition (ATR). This paper investigates the advantages / disadvantages of automatically guiding the selection of appropriate parameters for pulse coupled neural network factoring. The selected parameters would be targeted for use in assessing the state of the process being monitored. A noisy steady-state condition is to be distinguished from the approach of a target, or event of interest. Successful selection of appropriate parameters without expert intervention would increase the potential for automation of the system, and potentially simplify selection of the appropriate PCNN factor for analysis. The Powell method and evolutionary computation are compared and combined. The most promising options are recommended.

Original languageEnglish
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
PublisherSociety of Photo-Optical Instrumentation Engineers
Pages549-554
Number of pages6
Volume3718
Publication statusPublished - 1999
EventProceedings of the 1999 Automatic Target Recognition IX - Orlando, FL, USA
Duration: Apr 7 1999Apr 9 1999

Other

OtherProceedings of the 1999 Automatic Target Recognition IX
CityOrlando, FL, USA
Period4/7/994/9/99

Fingerprint

Neural networks
sensors
Sensors
pulses
Evolutionary algorithms
Automatic target recognition
target recognition
automation
Automation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Székely, G., Padgett, M. L., & Dozier, G. (1999). Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors. In Proceedings of SPIE - The International Society for Optical Engineering (Vol. 3718, pp. 549-554). Society of Photo-Optical Instrumentation Engineers.

Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors. / Székely, G.; Padgett, Mary Lou; Dozier, Gerry.

Proceedings of SPIE - The International Society for Optical Engineering. Vol. 3718 Society of Photo-Optical Instrumentation Engineers, 1999. p. 549-554.

Research output: Chapter in Book/Report/Conference proceedingChapter

Székely, G, Padgett, ML & Dozier, G 1999, Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors. in Proceedings of SPIE - The International Society for Optical Engineering. vol. 3718, Society of Photo-Optical Instrumentation Engineers, pp. 549-554, Proceedings of the 1999 Automatic Target Recognition IX, Orlando, FL, USA, 4/7/99.
Székely G, Padgett ML, Dozier G. Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors. In Proceedings of SPIE - The International Society for Optical Engineering. Vol. 3718. Society of Photo-Optical Instrumentation Engineers. 1999. p. 549-554
Székely, G. ; Padgett, Mary Lou ; Dozier, Gerry. / Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors. Proceedings of SPIE - The International Society for Optical Engineering. Vol. 3718 Society of Photo-Optical Instrumentation Engineers, 1999. pp. 549-554
@inbook{edc7552fffa04268b9fead7936657000,
title = "Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors",
abstract = "The focus of this paper is on real-world applications of neural networks and/or evolutionary computation, and on recent theoretical developments in parameter selection for pulse coupled neural networks (PCNN's) applicable to automatic target recognition (ATR). This paper investigates the advantages / disadvantages of automatically guiding the selection of appropriate parameters for pulse coupled neural network factoring. The selected parameters would be targeted for use in assessing the state of the process being monitored. A noisy steady-state condition is to be distinguished from the approach of a target, or event of interest. Successful selection of appropriate parameters without expert intervention would increase the potential for automation of the system, and potentially simplify selection of the appropriate PCNN factor for analysis. The Powell method and evolutionary computation are compared and combined. The most promising options are recommended.",
author = "G. Sz{\'e}kely and Padgett, {Mary Lou} and Gerry Dozier",
year = "1999",
language = "English",
volume = "3718",
pages = "549--554",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",
publisher = "Society of Photo-Optical Instrumentation Engineers",

}

TY - CHAP

T1 - Investigation of automated parameter adaptation in pulse coupled neural networks for spatially distributed sensors

AU - Székely, G.

AU - Padgett, Mary Lou

AU - Dozier, Gerry

PY - 1999

Y1 - 1999

N2 - The focus of this paper is on real-world applications of neural networks and/or evolutionary computation, and on recent theoretical developments in parameter selection for pulse coupled neural networks (PCNN's) applicable to automatic target recognition (ATR). This paper investigates the advantages / disadvantages of automatically guiding the selection of appropriate parameters for pulse coupled neural network factoring. The selected parameters would be targeted for use in assessing the state of the process being monitored. A noisy steady-state condition is to be distinguished from the approach of a target, or event of interest. Successful selection of appropriate parameters without expert intervention would increase the potential for automation of the system, and potentially simplify selection of the appropriate PCNN factor for analysis. The Powell method and evolutionary computation are compared and combined. The most promising options are recommended.

AB - The focus of this paper is on real-world applications of neural networks and/or evolutionary computation, and on recent theoretical developments in parameter selection for pulse coupled neural networks (PCNN's) applicable to automatic target recognition (ATR). This paper investigates the advantages / disadvantages of automatically guiding the selection of appropriate parameters for pulse coupled neural network factoring. The selected parameters would be targeted for use in assessing the state of the process being monitored. A noisy steady-state condition is to be distinguished from the approach of a target, or event of interest. Successful selection of appropriate parameters without expert intervention would increase the potential for automation of the system, and potentially simplify selection of the appropriate PCNN factor for analysis. The Powell method and evolutionary computation are compared and combined. The most promising options are recommended.

UR - http://www.scopus.com/inward/record.url?scp=0032642798&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032642798&partnerID=8YFLogxK

M3 - Chapter

AN - SCOPUS:0032642798

VL - 3718

SP - 549

EP - 554

BT - Proceedings of SPIE - The International Society for Optical Engineering

PB - Society of Photo-Optical Instrumentation Engineers

ER -