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

Metanatural Segmentation

IP.com Disclosure Number: IPCOM000076225D
Original Publication Date: 1972-Jan-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 2 page(s) / 40K

Publishing Venue

IBM

Related People

Baumgartner, RJ: AUTHOR

Abstract

Metanatural segmentation algorithms facilitate logical separation of touching characters in optical recognition systems, particularly where the system is required to recognize proportional-spaced characters. Generally, proportionally spaced characters touch each other, if at all, at relatively few points. The metanatural segmentation algorithms thus are slightly more powerful than natural, but less powerful than supernatural segmentation algorithms.

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 69% of the total text.

Page 1 of 2

Metanatural Segmentation

Metanatural segmentation algorithms facilitate logical separation of touching characters in optical recognition systems, particularly where the system is required to recognize proportional-spaced characters. Generally, proportionally spaced characters touch each other, if at all, at relatively few points. The metanatural segmentation algorithms thus are slightly more powerful than natural, but less powerful than supernatural segmentation algorithms.

The metanatural algorithms include ONE AND and NOT 3 SCAN AND. ONE AND requires not more than one pair of horizontally adjacent black bits in two adjacent scans. This is illustrated by the touching characters LA on document 5. As these characters are scanned by scanner 10, which includes cathode-ray tube 11, lens 12 and photomultiplier tube 13, the resultant video signal is quantized by cirmultiplier tube 13, the resultant video signal is quantized by circuitry 15. The quantized video is applied to shift register 20 and to segmentation logic circuitry 30 and 35. Logic 30 includes natural and supernatural segmentation circuitry, as described in U. S. Patent 3,526,876. Logic circuitry 35 implements the ONE AND and NOT 3 SCAN AND metanatural algorithms. The NOT 3 SCAN AND algorithm is satisfied when there is no occurrence of three horizontally adjacent bits in three adjacent scans, as illustrated by characters KO. The segmentation signals from 30 and 35 are applied to character recognition circui...