DYNAMIC PATTERN RECOGNITION BY MEANS OF THRESHOLD NETS,

Abstract

A method is expounded for the recognition of visual patterns. A circuit diagram of a device is described which is based on a multilayer threshold structure synthesized in accordance with the proposed method. Coded signals received each time an image is displayed are transmitted to the threshold circuit which distinguishes the signs, and from there to the layers of threshold resolving elements. The image at each layer is made to correspond with one resolving element the input of which receives the outputs from the resolving element that corresponds to the same image in the preceding layer, and also from the circuit that distinguishes the signs (characteristic features). As an example, the authors discuss the recognition of the ten numerals in their standard written form, constructed in a binary matrix 9x9 in size.

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1969
Accession Number
AD0697572

Entities

People

  • I. D. Ponomareva
  • V. A. Dolyatovskii

Organizations

  • National Air and Space Intelligence Center

Tags

DTIC Thesaurus Topics

  • Pattern Recognition
  • Recognition
  • Standards

Readers

  • Computer Programming and Software Development.
  • Image Processing and Computer Vision.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML