Browse Prior Art Database

Font Independent Character Recognition by Cryptanalysis

IP.com Disclosure Number: IPCOM000053032D
Original Publication Date: 1981-Aug-01
Included in the Prior Art Database: 2005-Feb-12
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Bahl, LR: AUTHOR [+2]

Abstract

A character recognition method is described for using the Baum-Petrie algorithm [2] to perform decipherment to thereby recover the original text.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 55% of the total text.

Page 1 of 1

Font Independent Character Recognition by Cryptanalysis

A character recognition method is described for using the Baum-Petrie algorithm [2] to perform decipherment to thereby recover the original text.

Ascher and Nagy [1] describe a method for data compaction for printed text digitized by an optical scanner. Compaction is achieved by transmitting or storing only the first instance of each pattern class and thereafter substituting this stored example for every subsequent occurrence of the symbol. No alphanumeric labels are attached to the pattern classes.

The method by which alphanumeric symbols could be attached to the pattern classes derived by the Ascher-Nagy method is based on the observation that if alphanumeric symbols are assigned arbitrarily to each pattern class (without repeating any symbol), the resultant string will be a simple substitution ciphertext version of the original text. By solving the ciphertext with use of cryptanalysis, the original text may then be recovered. A cryptanalysis technique for solving simple substitution ciphers is described in [2]. This method for cryptanalysis has been implemented in PL/I and is currently operational.

The overall character recognition method would then work as follows: 1. The document is scanned and digitized. 2. The scanned data is segmented into characters. 3. The characters are grouped into equivalence classes, as described in [1], or by some other clustering method. 4. Arbitrary symbols are assigned to e...