Browse Prior Art Database

Vertical Normalization Technique

IP.com Disclosure Number: IPCOM000077077D
Original Publication Date: 1972-Jun-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 2 page(s) / 39K

Publishing Venue

IBM

Related People

Glantz, WH: AUTHOR [+3]

Abstract

Shown is a normalization technique to measure and then adjust the height of characters to some standard height.

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 66% of the total text.

Page 1 of 2

Vertical Normalization Technique

Shown is a normalization technique to measure and then adjust the height of characters to some standard height.

Fig. 1 shows the prescan of a character to measure its height. During the measurement scans, the video is sampled once as each light-emitting diode (LED) is turned on in array scanner 1 (Fig. 3a).

The output of each scan is "ORed" with the previous scan and stored in profile register 2. At segmentation, the white bits in the bottom of the profile register 2 are counted to establish a reference between the bottom of the character and the bottom of the scan. Next, the black bits are counted to denote the number of bits or cells in the character height. These values are stored and used to normalize the character to some standard value during recognition scanning.

A basic clock system having a frequency ten times greater than the video sampling (Fig. 3a) provides a time base for normalization calculations. RH = CN = height of character in clock pulses where:

R = 10 (Number of pulses per sample during measurement scan)

H = Height of character in bits in measurement mode

N = Desired number of bits in normalization mode

C = Number of pulses per sample of video in normalization mode. For example, the number of bits counted during the measurement scan is sixteen (Fig. 3a). To normalize the character to a height of ten bits as shown in Fig. 3b, the number of bits is determined as follows: C = R over N . H = 10 over 10 . 16 = 16 pu...