Browse Prior Art Database

Print Recognition According to Character Asymmetry

IP.com Disclosure Number: IPCOM000096201D
Original Publication Date: 1963-Jan-01
Included in the Prior Art Database: 2005-Mar-07
Document File: 2 page(s) / 34K

Publishing Venue

IBM

Related People

Dersch, WC: AUTHOR

Abstract

This supplements or supplants conventional character recognition devices CR by the use of wave asymmetry detection devices in combination with memories. Two supplementary recognition techniques are combined. Thus, letters and words are classified into categories with high accuracy but poor resolution according to letter count LC and letter classification LCM analysis. They are then supplementarily analyzed with good resolution but poor accuracy by letter recognition devices LR and by the conventional character recognition devices CR.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

Print Recognition According to Character Asymmetry

This supplements or supplants conventional character recognition devices CR by the use of wave asymmetry detection devices in combination with memories. Two supplementary recognition techniques are combined. Thus, letters and words are classified into categories with high accuracy but poor resolution according to letter count LC and letter classification LCM analysis. They are then supplementarily analyzed with good resolution but poor accuracy by letter recognition devices LR and by the conventional character recognition devices CR.

The printed matter is scanned, character by character, by a pair of vertical SV and oblique SO scanners. The transduced optical outputs constitute signals to be analyzed in parallel and compared with code-word patterns. These are stored in a word-dictionary memory WD in which are placed code words according to letter count, initial letter type and terminal letter type.

The first classifier is letter counter LC where words are identified in the dictionary according to the number of letters they comprise. The second parameter measured is letter class so that words are classified according to the characteristics of their initial and terminal letters only. The characteristics detected here are letter asymmetry in both the vertical and oblique directions. The vertically scanned signal is differentiated at VD and then compared for asymmetry at detector VSA. The oblique is differentiated at OD and detected at OSA. A third analysis is performed by letter recognizer LR which, with rough accuracy, recognizes letters directly according to as...