A Method of Recognizing Single Strings of Nontouching Rotated Characters.

Abstract

This paper describes one method of recognizing single strings of nontouching, rotated characters of fixed-font. The method is based on a theory of two-dimensional moment invariants developed by M. K. Hu. Having input the visual image of a rotated word into the recognition system, each letter making up the string is first isolated in a systematic fashion. Then, twelve moment invariants, these components comprising a twelve-dimensional feature vector, are computed from each individual character image. Six of these features are based on the letter's silhouette and six on its boundary.

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1974
Accession Number
ADA004801

Entities

People

  • Thomas E. Southard

Organizations

  • Ohio State University

Tags

DTIC Thesaurus Topics

  • Behavior And Behavior Mechanisms
  • Boundaries
  • Geometry
  • Human Behavior
  • Identification
  • Mathematics
  • Personality
  • Recognition
  • Silhouettes
  • Two Dimensional

Readers

  • Computer Vision.
  • Speech Processing/Speech Recognition.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.