MXPA98006038A

MXPA98006038A - Detection to avoid falsification of currencies using lin detection

Info

Publication number: MXPA98006038A
Application number: MXPA/A/1998/006038A
Authority: MX
Inventors: Fan Zhigang; W Wu John; A Micco Felice; C Chen Mike; A Phong Kien
Original assignee: Xerox Corporation
Priority date: 1997-09-29
Filing date: 1998-07-27
Publication date: 1999-09-20

Abstract

The present invention relates to a detection method for preventing currency forgery, where the local edge information is used to accurately detect lines and curves of legitimate notes. The method can more accurately determine the location and orientation of a pattern or design and thus provides more reliable currency detection. An off-line detector is trained with exemplary notes which results in a stored pattern generated by registering a pattern or test design similar to a pattern to be tested, identifying the anchoring lines within the pattern, which are also represented in later test patterns. The pattern is rotated and displaced before matching it with the test pattern, so that the anchor lines align with the long lines detected within the test pattern. The pattern and the test pattern are then compared to determine if there is a similarity. The system comprises a microprocessor programmed to facilitate the training of an off-line detector with exemplary notes which were explored in the system where a pattern is generated by recording an image pattern of exemplary notes similar to a test pattern to be detected The microprocessor identifies the anchoring lines within the pattern, which are also presented in the test pattern, rotates and moves the pattern before matching it with the test pattern, so that the lines of enclamiento align with the lines that can be detected within the test pattern to determine if there are anchoring lines within the test pattern

Description

DETECTION TO AVOID FALSIFICATION OF CURRENCIES USING DETECTION OF LINES DESCRIPTION OF THE INVENTION Field of the Invention This invention relates generally to the techniques of image recognition based on a microprocessor and, more particularly, to a detection method to prevent counterfeiting of currency, which uses the detection of lines within of pictures.

Background of the Invention The ability to detect patterns or designs in an image may be useful in copying machines or scanning devices for the purpose of preventing counterfeiting. The challenge of incorporating such a method in the technology of current copiers or scanning devices is the difficulty in detecting images with rotation or displacement invariably. Specifically, the pattern or design could be in any orientation and anywhere in the image. The orientation and location of the currency or bank note can be relatively easy to estimate in the case of a single note REF. 27648 with a flat background; however, it becomes a major obstacle if multiple notes are involved and / or the notes are included in some background with a complicated image. Examples of identification and correction of the angle of inclination can be found in the following patents: U.S. Patent No. 5,528,387 to Kelly et al., Issued June 18, 1946, which teaches the registration of images in a scanning device. . In particular, the data of the edge of a document and the angle of inclination is calculated. The image is then rotated based on the angle of inclination and the non-image areas are filled using an image generation feature. U.S. Patent No. 4,922,350 to Rombola et al., Issued May 1, 1990, describes a two-step scanning apparatus for detecting the size and position of an original document on the plate of a scanning device. The signals of the image captured in a first scan are used to determine the limits and an amplification with the best adjustment, so that the image can be transferred to a record sheet, using the image signals generated in a scanning step later.

U.S. Patent No. 5,253,765 to Moorehead et al., Issued October 19, 1003, teaches a system for randomly classifying objects of different sizes (e.g. fungi). Variable moments are used, using the information of all the pixels for all the pixels within the limit of a captured image, to extract information about the size and orientation of the fungus. U.S. Patent No. 5,220,398 to Horn et al. Teaches an integrated analog VLSI chip that uses moments to determine the position and orientation of an object in a scene. In "Invariant Fitting of Planar Objets by Primitives", published in 1996 IEEE Proceedings of ICPR 96, pp. 508-512 Voss et al., Teaches a method of recognizing patterns or designs using primitive forms such as triangles, rectangles, circles, superquadratic elipses, etc. The author also describes a technique for describing primitive forms using moments in a normalized way; which results in a decrease in numerical work. In "Managing and Representing Image Workflow in Prepress Applications," Technical Association of the Graphic Arts (TAGA) Vol. 1, 1995 Proceedings, pp. 373-385, Venable et al., Teaches the use of structured images to handle the preprinted workflow. Detection methods to avoid previous falsifications depended on spot detection during recognition. As described in U.S. Patent No. 5,533,144 to Fan, entitled "Method and Detector of Patterns or Designs to Prevent Counterfeiting", a method and a detector are identified to prevent counterfeiting where, a flat image before being photocopied contains one or several patterns or bank note designs. With the method of? 144, the detection is invariably carried out with rotation and displacement. Specifically, the pattern or design can be in any orientation and anywhere in the image, and can be included in any background with complicated images. The image to be tested is processed block by block. Each block is examined to see if it contains a "walking point" by applying a detection and edge orientation estimation procedure. For a potential path point, an equalization procedure is performed against stored patterns or designs to decide whether the preset monetary note patterns or designs are valid once they are detected.

Other methods of pattern detection or designs to prevent counterfeiting are presented by the following patents: U.S. Patent No. 4,153,897 Yasuda, et al. Awarded on May 8, 1979 U.S. Patent No. 5,216,724 Suzuki, et al. Awarded on June 1, 1993 U.S. Patent No. 5,291,243 Heckman, et al. Awarded on March 1, 1994 Yasuda et al., Describe a system of recognition of patterns or designs in which the similarities between unknown and standard patterns are identified. The similarities are first detected in the respective displacement conditions where the unknown and standard patterns are relatively displaced from one another to the first limited degree, including the non-offset condition. Then the maximum value of these similarities is detected. The similarities are also detected in the respective displacement conditions where the unknown and standard patterns are relatively displaced from each other, a second degree 'greater than the first limited degree, where the condition of displacement giving the maximum value is that without relative displacement. Suzuki et al. Writes an apparatus for reading or processing images that can accurately identify a particular pattern, such as bank notes or securities. A detection unit detects the position information of an original image and a discrimination unit that extracts the pattern data of a certain part of the original image to discriminate whether the original image is the predetermined image based on the similarities between the data of the pattern and the default pattern. Heckman et al., Discloses a system for printing documents securely, which has copy detection or resistance to mishandling in a plurality of colors with a single-pass electronic printer. A validation signature has two half-tone patterns of intermixed color with half-tone density gradients that vary across the signature in opposite directions, but different from those in the background.

A fundamental problem in the recognition of patterns of electronic images is with the detection of lines, or more generally, the detection of curves. In pre-scanning an image the lines can be detected to decide if they are the edges of a paper document. Within the text of a document the lines can be used to estimate the angle of travel of the document. Also, the detection of lines or other simple curves often serves as the first step in the detection of a complex object, since the contour of a complex object can always be decomposed into simple curves. The detection of lines can be carried out by many standard methods such as that described by A. Rosefield and AC Kak in an article entitled "Digital Image Processing", Academic Press, Inc. In particular, this can be detected by the Hough Transformation, the which is probably the most important and most widely used line detection algorithm. This description employs the detection of lines to determine the location and orientation of the currency pattern in such a way that it has advantages over the previous detection methods. The proposed method is more accurate and requires fewer calculations.

All references cited here are incorporated by reference for their teachings.

Brief Description of the Invention To improve the speed and accuracy as described above, an electronic image detection method was used in which the information of the lines is used to determine the location and orientation of the currency presented. The detection method to avoid currency counterfeiting described uses the information of lines, instead of points, for detection to avoid falsification of notes or currencies. It is proposed the use of "walking lines" instead of "walking points" (as described in the * 144 of Fan) in the methods of currency image detection. An off-line detector is trained with, for example, notes that result in a pattern generated by the registration of a test pattern similar to a pattern to be tested; the walking lines within the pattern are identified. The lines are detected during the test; the pattern is rotated and shifted before matching the pattern to the test pattern so that the running lines are aligned along the lines detected within the test pattern; and the pattern and test pattern are compared to determine if there is a similarity.

The method can be carried out in a system comprising a microprocesprogrammed to carry out the above steps of the method. The microprocesof the system facilitates the training of an off-line detector with, for example, notes which are scanned in the system, where a pattern is generated by registering the pattern of an image of exemplary notes similar to a test pattern to be detected; identifies walking lines within the pattern. The lines are detected during the test; it rotates and shifts the pattern before matching the pattern to the test pattern, so that the running lines align along the lines that can be detected within the test pattern; and compares the pattern with the test pattern to determine if there are walking lines within the test pattern. The other advantages and salient features of the invention will become apparent from the detailed description which, taken in conjunction with the drawings, describes the preferred embodiments of the invention.

DESCRIPTION OF THE DRAWINGS Preferred embodiments and other aspects of the invention will become apparent from the following detailed description of the invention, when read in conjunction with the accompanying drawings, which are provided for the purpose of describing the embodiments of the invention. the invention and not to limit it, in which: Figure IA illustrates a test pattern or design that has walking (dotted) lines, which intersect in this sample - assuming an "L" shape; Figure IB illustrates the walking lines that can be found in a test image; Figure IC illustrates the detection of a legitimate currency after the pattern or design (Figure IA) is matched with the walking lines within the image (Figure IB); and Figure 2 is a block diagram of the system of the invention.

Detailed Description of the Invention To detect whether a copier was used to make counterfeits, a detector incorporated within the copier's scanning system is first trained off-line with exemplary notes. A pattern or design is created by recording the image pattern of the currency notes or target images (test image) to be , detected. The training includes generating the patterns and selecting one or several pairs of walking lines. The walking lines are straight, desirably long lines. They can be either straight lines within the currency's pattern, or edges of the currency. Each pair contains two lines which are orthogonal to each other in the direction. An off-line detector is trained with the images of the currency that result in a pattern generated by the registration of a pattern similar to the test pattern to be sampled; and identify the lines of walking within such pattern. In the detection, a Hough transformation, or any other standard methods, to detect long straight lines. The detected long lines are then grouped in pairs, each of which contains two long lines that are in orthogonal directions. Under the assumption that the pairs of long lines detected are the pairs of walking lines, the test pattern is matched with the patterns'. Figure IA illustrates a sample pattern. Specifically, the pattern is first rotated and moved before equaling, so that the running lines align with the detected long lines to be authenticated. This is best illustrated in Figure IB where the walking lines that are similar to the pattern are shown at the top of the Figure. A legitimate currency pattern to be detected after the pattern is declared (Figure IA) matched with the walking lines within the image (Figure IB), as shown in Figure IC, where the pattern (Figure IA), after of any necessary rotation and displacement, is superimposed on the upper part of the running lines of the test pattern of Figure IB. The results of the test can then be determined as positive. Since there are fewer "long lines" in an image than in blocks containing points or edges, the number of pattern equalization routines required for the determination of similarity is less in the proposed method than in that of the Fan x144. Also, since long lines contain more information about the direction of the line than short edges, the estimation of the orientation is much more accurate. As a result, the reliability of equalization of the pattern or design is also improved. Referring to Figure 2, the system equipment needed for the detection system to prevent counterfeiting 5 could include a scanning device 1 for receiving the test image, a microprocessor 2 programmed to detect counterfeit currencies by detecting lines in the scanned documents, a memory 3 for storing the test patterns, and indicator means 4 for indicating the detection of the counterfeit and being able to further avoid the duplication of test patterns that may be found in a larger system (not shown) such as a copier or of a computer. The indicating means 4 could provide an output signal to the larger systems indicating the state of the tested image. The detection system to avoid counterfeiting could facilitate the training of the microprocessor 2 (also known as "detector") offline with notes or currencies, which are scanned 1 in the system where a pattern is generated by the registration 3 of a pattern of images of exemplary notes similar to a test pattern to be detected. The running lines within the pattern, which are additionally represented in the test pattern, could be identified by the microprocessor 2. The microprocessor system could perform the detection of lines during the test. It will also allow the rotation and displacement of the pattern before matching it with the test pattern, so that the running lines can be aligned with the long lines that can be detected within the test pattern. The system 2 could then compare the pattern, stored in memory 3, with the test pattern that has been scanned 1 or seen by means of image capture to determine if there are walking lines within the test pattern.

Although the invention has been described with reference to a particular embodiment, it is intended that this particular embodiment be illustrative, not limiting. Various modifications can be made without departing from the spirit and scope of the invention as defined in the amended claims. Modifications and alterations will occur to anyone after reading and understanding this specification; therefore, it is intended that all such modifications and alterations be included as much as possible within the scope of the appended claims or equivalents thereof.

It is noted that in relation to this date, the best method known to the applicant to carry out the aforementioned invention, is that which is clear from the present description of the invention. Having described the invention as above, property is claimed as contained in the following:

Claims

1. A false currency detection method, characterized in that line information is used to determine the location and orientation of a pattern or image design and the pattern is then authenticated by matching such pattern or design with a pattern or design.

The method according to claim 1, characterized in that an off-line detector is trained with exemplary notes that results in at least one pattern or design generated by registering a test pattern similar to a pattern or design to be tried; the walking lines within such pattern are identified, which are further represented in the test pattern; the pattern is rotated and moved before the pattern is squared with the test pattern, so that the running lines are aligned with the long lines detected within the test pattern; and the pattern and test pattern is compared to determine if there is a similarity.

3. A method for detecting false currencies using the line detection information, characterized in that it comprises: training an off-line detector with exemplary notes, where at least one pattern or design is generated by recording a pattern of note images specimens similar to a test pattern to be detected; identify walking lines within the pattern, which are further represented in the test pattern; rotate and move the pattern before matching the pattern with the test pattern, so that the running lines align with the long lines that may be detected within the test pattern; and compare the pattern with the test pattern to determine if there are walking lines within the test pattern

4. The method in accordance with the claim 3, characterized in that the training further comprises generating patterns by selecting at least a pair of walking lines within the exemplary notes, which may be, for example, straight lines within a currency or the edges of a currency, and where at least a pair of walking lines contain two lines which are orthogonal to each other in the direction.

5. The method of compliance with the claim 4, characterized in that one or more pairs of running lines can comprise straight lines within the design pattern of a currency.

6. The method according to claim 4, characterized in that one or several pairs of running lines may comprise straight lines representing edges of the currency within a pattern or design of a currency.

7. A system for detecting counterfeits, characterized in that it comprises a microprocessor programmed to use known information of local edges within authentic currency documents that are scanned in such a system to detect running lines within authentic currency documents, wherein the lines are used to determine the existence of them within a tested document, where the location and orientation of the pattern of a pattern image derived from the images of the currency stored in the memory can be compared with the test document, after that the path lines are detected within the test document and aligned with the path lines of the authentic currency document, matching the test document with the pattern of the standard image to determine the authenticity of the test document.

The system according to claim 7, characterized in that it also comprises indicating means to indicate if the test pattern is false based on the comparison.

The system according to claim 8, characterized in that the indicating means provide inputs to a controller that avoids photocopying or storing the test pattern.

10. The system in accordance with claim 8, characterized in that it also comprises a memory for storing such patterns or designs.

The system according to claim 8, characterized in that it further comprises a scanning device for scanning images in the system.

12. A counterfeit detection system, characterized in that it comprises a microprocessor programmed to: facilitate the training of an off-line detector with authentic currency documents, which are scanned in the system, where a plurality of patterns or designs are generated. registering the pattern of an image within the authentic currency documents, resulting in at least one pattern or design where at least one pattern is stored in a memory; identify walking lines within at least one pattern or design which are only represented in authentic currency documents; rotate and move at least one pattern during counterfeit detection operations, where the equalization of the reindeer a pattern or as a scanned test pattern is facilitated, so that the walking lines aligned with similar lines can be detected within the test pattern, thus resulting in an alignment; and comparing at least one pattern with the test pattern to determine if there are walking lines within such a pattern and nrueba.

The system according to claim 12, characterized in that it also comprises a memory for storing patterns or designs.

The system according to claim 12, characterized in that it further comprises a scanning device for scanning images in the system.

15. The system according to claim 12, characterized in that it also comprises indicating means to indicate if the test pattern is falsified based on the comparison.

16. The system according to claim 15, characterized in that the indicating means provide inputs to a controller that prevent the photocopying or storage of the test pattern.

The system according to claim 13, characterized in that furthermore a scanning device for scanning images in the system.

18. The system according to claim 17, characterized in that it also comprises indicating means to indicate if the test pattern is falsified based on the comparison.

19. The system according to claim 18, characterized in that the indicating means provide inputs to a controller that prevent the photocopying or storage of the test pattern. The system according to claim 7, characterized in that it further comprises: indicating means for indicating if the test pattern is false based on the comparison, wherein the indicating means provide inputs to a controller that prevent photocopying or storage of the test pattern; a memory to store the patterns; and a scanning device to scan images in the system. SUMMARY OF THE INVENTION A detection method is described to avoid counterfeiting currencies, where local edge information is used to accurately detect lines and curves of legitimate notes. The method can more accurately determine the location and orientation of a pattern or design and thus provides more reliable currency detection. An off-line detector is trained with exemplary notes which results in a stored pattern generated by the registration of a pattern or test design similar to a pattern to be tested; the walking lines within the pattern are identified, which are also represented in later test patterns. The pattern is rotated and moved before matching it with the test pattern, so that the running lines align with the long lines detected within the test pattern. The pattern and test pattern are then compared to determine if there is a similarity. The system comprises a microprocessor programmed to facilitate the training of an off-line detector with exemplary notes which were explored in the system where a pattern is generated by recording an image pattern of exemplary notes similar to a test pattern to be detected The microprocessor identifies the running lines within the pattern, which are also presented in the test pattern; rotates and moves the pattern before matching it with the test pattern, so that the running lines align with the lines that can be detected within the test pattern; and compares the pattern with the test pattern to determine if there are walking lines within the test pattern.