PERCEIVING AND RECOGNITION AUTOMATA

Abstract

The utility of the perceptron in recognizing imperfect, mixed-font printed characters was investigated. A general-purpose computer implementation of perceptrons which incorporates a high degree of flexibility was used. The ability of such an organism to recognize a complete alphabet characterized by a limited amount of distortion was demonstrated. The same machine was used to establish that recognition of limited sets of the alphabet could be accomplished usingA WIDE DEGREE OF DISTORTION, TRANSLATION, AND NOISE. (Author)

Document Details

Document Type
Technical Report
Publication Date
Dec 31, 1961
Accession Number
AD0272245

Entities

People

  • H.r. Leland

Organizations

  • Calspan

Tags

DTIC Thesaurus Topics

  • Alphabets
  • Automata
  • Computers
  • Distortion
  • Machines
  • Personality
  • Recognition
  • Resilience
  • Translations

Readers

  • Computational Linguistics
  • Neural Network Machine Learning.
  • Systems Analysis and Design