Browse Prior Art Database

Character Separation for Proportional Space Printing

IP.com Disclosure Number: IPCOM000073442D
Original Publication Date: 1970-Dec-01
Included in the Prior Art Database: 2005-Feb-22
Document File: 2 page(s) / 43K

Publishing Venue

IBM

Related People

McCullough, JW: AUTHOR

Abstract

Many techniques exist for the segmentation of poorly separated characters for optical character-recognition machines. Many of those, however, are so strong that they may break single character images into many separate parts, if allowed to run at all times during the scanning of a character. Accordingly, "sectioning" has been developed to enable certain segmentation schemes only in certain areas of each character. In view A, video detector 10 reads images into storage 11, which are then recognized by logics 12. The end of each image to be recognized is determined by a signal on line 13 from a number of segmentation generators 14.

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 54% of the total text.

Page 1 of 2

Character Separation for Proportional Space Printing

Many techniques exist for the segmentation of poorly separated characters for optical character-recognition machines. Many of those, however, are so strong that they may break single character images into many separate parts, if allowed to run at all times during the scanning of a character. Accordingly, "sectioning" has been developed to enable certain segmentation schemes only in certain areas of each character. In view A, video detector 10 reads images into storage 11, which are then recognized by logics 12. The end of each image to be recognized is determined by a signal on line 13 from a number of segmentation generators 14. Generators 14 are enabled (individually or simultaneously) by section computer 15, which is reset by segmentation signal 13, and which computes a function of certain topological pattern features or measurements 16 derived from logical functions of the image bits in storage 11.

The measurements 16 taken for each scan n are the number of leading-edge bits Pn, the number of trailing-edge bite Nn, and the bit density Ln. These may be combined in computer 15 to generate, to each scan, a threshold function of the form

(Image Omitted)

where some or all of U, V, W, X and Y may be functions of the number n of scans since the preceding segmentation signal on line 13. In a representative embodiment, U and Y are constants, while V, W and X are functions of a feedback factor F, which depends upon n...