Browse Prior Art Database

Magnetic Image Processing

IP.com Disclosure Number: IPCOM000121496D
Original Publication Date: 1991-Sep-01
Included in the Prior Art Database: 2005-Apr-03
Document File: 2 page(s) / 80K

Publishing Venue

IBM

Related People

Rodite, RR: AUTHOR

Abstract

Disclosed is a system and method for enhancing the accuracy of character recognition performed on certain types of documents. As applied to processing the amounts written on bank checks, the disclosed method eliminates the preprinted areas of the check from the recognition process. This removes any uncertainty in the recognition process attributable to indicia which do not need to be recognized.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Magnetic Image Processing

      Disclosed is a system and method for enhancing the
accuracy of character recognition performed on certain types of
documents.  As applied to processing the amounts written on bank
checks, the disclosed method eliminates the preprinted areas of the
check from the recognition process. This removes any uncertainty in
the recognition process attributable to indicia which do not need to
be recognized.

      According to Fig. 1, the check 10 is moved in direction 12 by
the transport conveyor 14. This movement allows the optical image
scanner 16 to obtain data representing the optical indicia present on
the check. The optical indicia include the preprinted data and marks
on the check, plus the information added by the person or machine
completing the check. The check is then moved underneath the magnetic
image scanner 18 to obtain data representing the magnetic indicia
present on the check. The scanner 18 can be similar to the
conventional MICR devices which read preprinted magnetic codelines
usually present on checks.  The preprinted high-contrast areas of the
check use the same type of magnetic ink or toner as customarily used
for the codeline. Since the written information on the check is not
in magnetic ink, the data from scanner 18 does not include the
courtesy amount placed on the check.

      The data from the two scanners is applied to the image
processor 20 where the magnetic image data is subtracted from the
optical image data. The resulting data, therefore, represents the
information added to the check, and this information is applied to
the character recognizer 22. Since this data represents less
information than obtained by the optical scanner 16, detecting
character patterns in the data is easier.  The subtraction of the
da...