Feature selection for PCNN exemplars as applied to electronic nose analysis and ATR

M. L. Padgett, T. A. Roppel, V. Becanovic, Geza Szekely, J. Waldemark

Research output: Contribution to journalConference article

Abstract

A novel methodology is suggested for the selection of exemplars in algorithm training and testing. The practicality of this new approach is illustrated in relation to Automatic Target Recognition (ATR) by an electronic nose system. The proposed method, combining Pulse Coupled Neural Networks Factoring with Evolutionary Algorithms is compared with use of a traditional Principal Component Analysis, and a development strategy combining the two is suggested.

Original languageEnglish
Pages (from-to)542-548
Number of pages7
JournalProceedings of SPIE - The International Society for Optical Engineering
Volume3718
Publication statusPublished - Jan 1 1999
EventProceedings of the 1999 Automatic Target Recognition IX - Orlando, FL, USA
Duration: Apr 7 1999Apr 9 1999

    Fingerprint

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this