Browse Prior Art Database

Software Normalization of Numeric Hand Printed Characters

IP.com Disclosure Number: IPCOM000079120D
Original Publication Date: 1973-May-01
Included in the Prior Art Database: 2005-Feb-26
Document File: 3 page(s) / 141K

Publishing Venue

IBM

Related People

Glantz, WH: AUTHOR [+2]

Abstract

Numeric hand-printed characters vary significantly in size and position so that a vertical raster sufficient to guarantee reading most of the characters, often includes fragments of characters located above and below the scan line. The above flow chart describes an algorithm for excluding these fragments when measuring characters for normalization.

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 3

Software Normalization of Numeric Hand Printed Characters

Numeric hand-printed characters vary significantly in size and position so that a vertical raster sufficient to guarantee reading most of the characters, often includes fragments of characters located above and below the scan line. The above flow chart describes an algorithm for excluding these fragments when measuring characters for normalization.

The raster is divided into three zones, each of which can be treated differently. The high zone is most likely to contain fragments from above the line of scan, while the low scan is most likely to contain fragments below the line of scan. The central zone normally contains the character of interest, although this character may extend into either or both of the other zones. As the character is scanned, a profile is constructed by ORing together all scans. The height of the character is then determined by examining the profile.

Referring to the flow chart, characters first encountered and ending in the central zone are assumed to be correctly positioned. If the central minimum character requirement (CMCR) is satisfied, a C width counter is started and all subsequent scans are ORed together. The address of the CMCR is stored in memory. At segmentation, tests are made to ensure there is no connecting video between the central zone and either the high or low zones. If all tests prove positive, the normalization routine normalizes the character and video is sent to the recognition circuitry.

In the case where video is first encountered in the high zone, a test is made to determine that a high minimum character requirement...