Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Virtual Normalization for Character Recognition

IP.com Disclosure Number: IPCOM000044762D
Original Publication Date: 1984-Jul-01
Included in the Prior Art Database: 2005-Feb-06
Document File: 1 page(s) / 11K

Publishing Venue

IBM

Related People

Bednar, GM: AUTHOR [+2]

Abstract

Variable size character images are normalized to fit a standard picture element matrix by a virtual, rather than a real transformation, to present image data of a standardized size to character recognition logic.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 1

Virtual Normalization for Character Recognition

Variable size character images are normalized to fit a standard picture element matrix by a virtual, rather than a real transformation, to present image data of a standardized size to character recognition logic.

Virtual normalization is accomplished by computing the individual bit addresses in the unnormalized image which are to be selected as image bits for recognition. One known normalization technique selectively deletes rows and columns of image bits. This technique is implemented virtually by computing the vertical and horizontal offsets 10 and 11 from a starting address as a function of the size of the unnormalized character image. Thus, for a character 12 whose unnormalized image size is 18 x 26 to be virtually normalized to a standard 7 x 9 format, offsets from a starting element 13 would be computed as shown in the Figure.

The recognition logic tests only the bits of the stored image that correspond to the addresses of the virtual image bits, as indicated by the circles 14.

Disclosed anonymously

1