Image analysis – Applications – Reading paper currency
Reexamination Certificate
1999-10-28
2003-04-22
Boudreau, Leo (Department: 2621)
Image analysis
Applications
Reading paper currency
C382S209000, C382S278000
Reexamination Certificate
active
06553136
ABSTRACT:
TECHNICAL FIELD
The invention relates generally to image, processing and more particularly to a system and a method for detecting images of restricted documents for counterfeit protection.
BACKGROUND ART
Counterfeiting of restricted documents, such as monetary bills, is a wide-spread problem throughout the world. Every year, an unknown amount of counterfeit currency is introduced into our society. Historically, production of counterfeit currency required an expert artist to engrave the intricate details of paper money on plates. As such, counterfeiting was limited to a relatively small number of criminals. However, with the advance of color scanning and printing technologies, more people are capable of counterfeiting restricted documents, such as monetary bills, using copy machines and scanners. As copy machines and scanners become more sophisticated and less expensive, a serious threat exists that these machines will be used for counterfeiting purposes by people who might otherwise have had no means to make counterfeit documents.
In response to this threat, image processing systems have been developed that can be incorporated into a copy machine to thwart counterfeiting activities. These image processing systems operate to detect counterfeiting activities by recognizing that a restricted document is being copied using the copy machine. Recognition is accomplished by comparing an entire scanned image with entire images of restricted documents, typically, by performing straightforward linear convolutions. If the scanned image is determined to match one of the restricted-document images, the copy machine is prevented from printing a hard copy of the scanned image. Examples of such image processing systems are described in U.S. Pat. No. 5,363,202 to Udagawa et al. and U.S. Pat. No. 5,659,628 to Tachikawa et al.
A concern with the use of a known image processing system to detect counterfeiting activities is that a large amount of memory space is required to store the images of the restricted images. If the image processing system is to recognize any valid paper money of a designated country, both front and back images of all the paper money for that country would need to be stored in the copy machine in order for the image processing system to make the appropriate comparison between the scanned image and the images of the restricted documents. Although the required images that need to be stored is limited in number, these images are stored in high resolution for reliable image comparison, which takes up sufficient amount of memory. Another concern is that due to the intense processing involved in making the comparisons, the execution time of the image processing for a single scanned image may be too long for any practical implementation.
In light of these concerns, what is needed is a system and a method of detecting counterfeiting activities, with the system and method being efficient with respect to detection speed and memory requirement.
SUMMARY OF THE INVENTION
A system and a method for preventing counterfeiting activities utilize an image detection scheme that uses only a few portions of restricted-document images, i.e., templates, to determine whether an input image is a duplicate of a restricted-document image. The image detection scheme can reliably operate at very low resolutions, which allows the use of small kernels for image correlation and reduces the number of pixels that have to be actually processed. The system is designed to be incorporated into a host device, such as a copy machine, a scanner, a personal computer or a printer, to detect a potential counterfeiting activity, so that an appropriate action can be taken to prevent and/or to report the detected activity. The image detection scheme allows the system to operate at an increased detection speed, while minimizing the memory requirement.
The counterfeit protection system includes a training unit and an image detection unit. The training unit is configured to execute an “off-line” training procedure, during which a parameter file of one or more input restricted-document images is generated. The parameter file contains detection descriptors for the input restricted-document images. During an “on-line” image detection procedure, the image detection unit of the counterfeit detection system utilizes the parameter file to determine whether a given image closely resembles one of the input restricted-document images. In a preferred embodiment, the training unit is a separate unit that is not included in the host device.
The training unit of the counterfeit protection system includes a training module and a storage unit. The storage unit may be a conventional hard disk drive. The training module of the training unit includes a kernel-template selector and a template parameter computer. The kernel-template selector is comprised of a kernel selector and a template selector. The kernel selector is configured to select kernels that are either universal or customized. Similarly, the template selector is configured to extract templates from the input restricted-document images that are either kernel-dependent or kernel-independent. In the preferred mode of operation, universal kernels and kernel-dependent templates are selected to reduce the memory requirement, while also minimizing false alarms. After the kernels and templates have been selected, the template parameter computer projects the templates onto the kernels to derive projection values. The derived projection values are used by the template parameter computer to define a number of intervals that will be used by the image detection unit during the on-line detection procedure. The template parameter computer then stores the defined intervals in the storage unit of the training unit. The defined intervals are included in the parameter file. The parameter file also includes a number of input parameters, the coefficients of the selected kernels, and document parameters.
The image detection unit of the counterfeit protection system includes a detection module and a signaling unit. The detection module of the image detection unit includes a candidate selector and a geometric verifier. The candidate selector is configured to receive an input image and to access the parameter file during an on-line detection procedure. In one embodiment, the counterfeit protection system includes a read-only storage device that may be either replaceable or non-replaceable. In this embodiment, the parameter file generated by the training unit is stored in the read-only storage device. If the read-only storage device is replaceable, the parameter files may be periodically updated by replacing the storage medium utilized by the read-only storage device. However, the counterfeit protection system may be configured to operate without the read-only storage device. In this configuration, the parameter file may be uploaded to a read/writeable memory of the image detection unit from an external source, such as the Internet.
The candidate selector of the image detection unit is comprised of a linear match filter and a thresholder. The linear match filter is designed to project portions of the input image onto the kernels that were selected during the off-line training procedure. The resulting projection values are then compared using the defined intervals of the parameter file by the thresholder. Portions of the input image that satisfy the comparison for all intervals are declared template match candidates. Coordinates of these template match candidates, as well as corresponding restricted-document image indices, template numbers and estimated orientation angle intervals, are transmitted to the geometric verifier. The geometric verifier is configured to compare the relative coordinates of the template match candidates in the input image with the relative coordinates of the templates in the input restricted-document images to determine whether the input image is a duplicate of a restricted-document image. If the relative positions of the template match candidates are similar to those of the templa
Keshet Renato
Shaked Doron
Akhavannik Hussein
Boudreau Leo
LandOfFree
System and method for counterfeit protection does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for counterfeit protection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for counterfeit protection will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3062342