Optical implementations of radial basis classifiers

Mark A Neifeld, Demetri Psaltis

Research output: Contribution to journalArticle

16 Scopus citations


This paper describes an optical system architecture for implementation of a radial basis function (RBF) pattern classifier, and another optical disk based system for handwritten character recognition. A processing rate of 2,600,000 binary operations/s, equivalent to 13,000 Euclidean distances/s is achieved, limited by an upper processing boundary of nearly 1011 binary operations. Another system that demonstrates additional robustness and on-line learning capability is also discussed.

Original languageEnglish (US)
Pages (from-to)1370-1379
Number of pages10
JournalApplied Optics
Issue number8
Publication statusPublished - Jan 1 1993


ASJC Scopus subject areas

  • Atomic and Molecular Physics, and Optics

Cite this