Browse Prior Art Database

Apparatus for OCR Postal Address Code Determination

IP.com Disclosure Number: IPCOM000082045D
Original Publication Date: 1974-Sep-01
Included in the Prior Art Database: 2005-Feb-28
Document File: 3 page(s) / 53K

Publishing Venue

IBM

Related People

Ett, AH: AUTHOR [+2]

Abstract

Described is apparatus for determining the correct characters provided. as an optical character recognition (OCR) output when used to read alphameric or numeric codes of a structured nature, such as used in a postal address code (PAC).

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 53% of the total text.

Page 1 of 3

Apparatus for OCR Postal Address Code Determination

Described is apparatus for determining the correct characters provided. as an optical character recognition (OCR) output when used to read alphameric or numeric codes of a structured nature, such as used in a postal address code (PAC).

The method employed involves the use of multiple OCR outputs representing possible recognition characters above a set recognition threshold, combined with the feature vector used in the recognition process. From these data a character likelihood probability index is extracted from a read-only store, and then combined with the binary character representation. This output for each character possibility is formatted and stored in a shift register matrix.

Each set of possible characters is analyzed on the basis of its likelihood probability, convolved with the a priori probability of its being in that position in the code considering all previous characters defined. The highest probability is then selected and the character selected transferred to the output register.

The implementation involves an OCR 1 which provides a set of outputs of possible recognition characters which exceed a preset recognition threshold criteria, and the feature vector used to obtain these output characters. Each of these possible output characters is sequenced in order of probability according to the AOCR logic, to a read-only store (ROS) 3 by the sequencer 2. The read-only store is addressed by the feature vector and the character code.

At each address of the ROS 3 is stored a word providing the character code, and probability that the character feature vector represents that character. Operation of the ROS and other elements of the system are controlled by the clock and timing chain 4. The output of the ROS is p...