Image analysis – Image sensing – Multiple scanning
Reexamination Certificate
1998-08-19
2001-06-05
Mancuso, Joseph (Department: 2621)
Image analysis
Image sensing
Multiple scanning
C382S320000
Reexamination Certificate
active
06243504
ABSTRACT:
TECHNICAL FIELD
The present invention relates in general to a system and method for performing character recognition of printed media, and more particularly, to an integrated system and method employing optical and magnetic ink character recognition for providing identification and processing of printed media having magnetically encoded characters.
BACKGROUND INFORMATION
Various applications require the automated processing of materials having printed characters. Particularly for applications which are perfunctory in nature such as bill statement processing, payment check processing, or account credit processing, a system and method which can provide for high-accuracy, automated processing requiring little or no human intervention would be desirable for a number of factors. One very practical motivation for the development of such a system and method is the reduction of man-hours required to perform these tasks. Such a reduction in man-hours would correspondingly result in cost savings and increase efficiency. Another inherent advantage provided by an automated, high-accuracy system and method is a reduction in human error thereby providing more reliability and efficiency.
One particular application which requires the processing of printed media that is well-known to a vast majority of the population is the processing of a payment check at a point of sale within the context of a financial transaction. A typical point of sale system might constitute a cash register at a checkout counter within a commercial retail store. In such an instance, either one or very few employees perform all of the duties associated with receiving payment for the purchased items, packaging the purchased items, and providing the customer with a sales receipt. If an automated system and method were developed which would perform one or more of the tasks required of the employee at the point of sale, then the employee would have more free time and could then provide the customer with faster and more efficient service.
One such point of sale system has been developed in an attempt to assist the processing of payment checks within the context of a financial transaction. Part of the check processing includes reading of special characters located at the bottom of each check which are called Magnetic Ink Character Recognition (MICR) characters. These characters were developed by and are used by the banking industry to assist in payment check processing and tracking once a payment check leaves a retail store. The MICR characters include such information as bank routing number, customer account number, check sequence number, or other information as specified by individual banks or credit institutions. MICR characters are specially formed characters that are similar to normal digits of zero to nine. Additionally, they include an additional four special characters that are used to separate the various MICR fields. The MICR characters are printed with a magnetic ink such that they can be magnetized and read “magnetically.”
Reading characters magnetically using a magnetic ink character recognition system offers a certain degree of protection against fraudulent activities including the alteration or modification of the characters' appearance. Consumer stores that accept payment checks will often use the MICR information to verify that the check is valid and acceptable for tender. A typical store solution for performing payment check processing will often include input/output (I/O) devices such as printers, keyboards, displays, and MICR readers. One disadvantage of using a separate MICR reader is that it requires additional counter space, communications cables, and power cables. Kinney, et al., U.S. Pat. No. 5,613,783 addresses the integration of a MICR reader within the confines of the printer housing for maximizing counter space usage of a point of sale system.
Magnetically reading MICR characters typically utilizes a read head having a single channel that is passed over the MICR character string (sometimes referred to as a MICR codeline). Signals are produced when the read head traverses from a non-magnetic region (paper background) to a magnetic background (edges of a MICR character) or from a magnetic region to a non-magnetic region. Each time this boundary is passed a pulse is generated in the read head. A read head coming into a character edge will produce a positive signal whereas a read head coming out of a character edge will produce a negative signal. These signal images are converted into digital data, stored, and then compared to known images of MICR characters using a magnetic image processor comprising a computer microprocessor. Typically, the comparisons are based upon number and type of pulses (positive or negative), time between pulses, and the magnitude of each pulse. Once recognized by the decode algorithm the data is passed and displayed using a data transmission and display system for further processing. Such a system might typically include data transmission cables and a video display monitor for displaying relevant information concerning the printed media having the magnetic ink character string.
In reality, the formed MICR characters of a printed medium often possess irregularities which make difficult the proper recognition of the magnetic ink character string. The magnetic ink characters might not be perfect in size, location on the check, or magnetic particle density. Additionally, the magnetic ink character string might possess inking problems making even more difficult proper recognition. Any of these deleterious characteristics may create problems for a decode algorithm tasked with performing character recognition resulting in either a non-readable character or incorrect reading of a character, commonly referred to as a “substitution error.” Furthermore, given that this character recognition technology relies solely on magnetic mechanisms, external magnetic sources may create noise problems of sufficient magnitude as to cause errors in recognizing the magnetic ink character string.
“Read rate” is a term of the character recognition art used to describe the percentage of error free reads which are performed in a given group of character recognition attempts. Typical read rates for magnetic ink character recognition systems are in the mid-ninety percent range. With “clean checks” that do not contain any of the inking problems previously described, the read rates may be as high as one hundred percent whereas checks that do contain inking problems typically have read rates near ninety percent. A mix of typical checks which passes through a point of sale, payment check processing system will commonly have a read rate in the mid-ninety percent range.
Another problematic source that may contribute to low read rates is magnetic or electrical field noise, commonly referred to as electromagnetic interference. This type of noise may typically be generated from a display system or a security tower, both of which are common devices in a point of sale system. In some instances, the electronic devices contributing to these noise sources can be physically moved resulting in a lowering of the electromagnetic induced noise from the emitted field of the noise source. However, in many cases these devices may not be moved as they are required to be in close proximity of a point of sale system for providing services to assist in the financial transaction. A monitor will produce a repeatable noise pulse which may be coupled magnetically into the coil of the read head. This noise source can affect the ability of the MICR algorithm to properly and accurately decode the magnetic ink character string.
Changes in the banking industry may render typical read rates in prior art MICR systems inadequate. The banking industry has been considering a new process wherein a check would be read at a point of sale station system, the account of the payment check adjusted, the front and back sides of the payment check printed with the appropriate transaction information, and the completely processed payment check then re
Bayat Ali
International Business Machines - Corporation
Mancuso Joseph
Schelkopf J Bruce
Winstead Sechrest & Minick P.C.
LandOfFree
Integrated magnetic ink character recognition system and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Integrated magnetic ink character recognition system and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Integrated magnetic ink character recognition system and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2474237