AU2009238260A1 - Forgery detection using finger print - Google Patents

Description

S&F Ref: 920310 AUSTRALIA PATENTS ACT 1990 COMPLETE SPECIFICATION FOR A STANDARD PATENT Name and Address Canon Kabushiki Kaisha, of 30-2, Shimomaruko 3 of Applicant: chome, Ohta-ku, Tokyo, 146, Japan Actual Inventor(s): Amit Kumar Gupta Stuart William Perry Address for Service: Spruson & Ferguson St Martins Tower Level 35 31 Market Street Sydney NSW 2000 (CCN 3710000177) Invention Title: Forgery detection using paper finger print The following statement is a full description of this invention, including the best method of performing it known to me/us: 5845c(2389567_1) -1 FORGERY DETECTION USING PAPER FINGER PRINT TECHNICAL FIELD The current invention relates generally to document security and in particular to 5 detecting forged documents. BACKGROUND The ability to establish the authenticity of a physical medium is a critical requirement of many businesses. Also, in cases where a physical medium represents a 10 significant monetary or societal value, or where a reasonable facsimile of an object can be created at a much lower cost than the cost to acquire the original, criminals will be tempted to create a forgery and misrepresent the forgery as the authentic object. In particular, in the case of paper-based physical media such as administrative certificates, licenses, contracts, company letterhead paper and other related examples, possession of 15 the original document may represent substantial value to the bearer. Techniques exist to facilitate document authentication. A substantial set of these techniques rely on a physical addition to a document, for example in the form of a bar code, a hologram or a watermark. However, sophisticated counterfeiting techniques often eventually defeat these methods. The reason for this is that since the creation of the 20 physical addition is not out of reach of the object's manufacturer, as counterfeiting techniques become more sophisticated, the creation of the physical addition eventually becomes within reach of the counterfeiter. Additionally, physical additions to the original document may not be preferred due to the costs associated with the addition. To address these disadvantages, recently developed authentication systems rely on a randomly 25 varying paper fibre structure, which is as unique to each paper sample as a human finger print is unique to a human being. Some existing document fingerprinting methods exploit the unique speckle pattern produced by shining coherent light onto a piece of paper, allowing a test to be developed for matching a strip of paper with its previously-scanned "fingerprint". Other 30 techniques have achieved similar results at lower cost without requiring the use of 920310 (2384075_1) -2 coherent light, making it possible to authenticate a paper fingerprint created using a standard scanner or photocopier. Authentication systems based on a standard scanner have an advantage of not requiring sophisticated equipment, and hence more commercial applications are possible. 5 In some cases, a counterfeiter may be able to defeat a paper-fingerprint-based authentication system using an over-print method. An over-print method is a method by which pattern matching a previously scanned pattern is printed on the paper. That is, a 2D image is printed on the paper representing a facsimile of the paper fingerprint. Notionally, this involves printing a 2D representation of the 3D structure of the original 10 paper on the paper of the forgery. The printing equipment could be a standard printer or a sophisticated system such as laser engraving. For example, a laser printer may print toner on the forgery paper, masking the underlying structure of the forgery paper, to reproduce the paper fingerprint. Using a standard printer for an over-print attack may allow the generation of a counterfeit that appears authentic to a paper-fingerprint-based 15 authentication system using standard equipment. The resistance of the authentication system can be increased using additional security layers, such as public/private key for encrypting the paper fingerprint. However adding additional layers of security to the paper fingerprint system does not directly address the problem of overprint attacks. If the key is compromised, the authentication 20 system is just as vulnerable to an overprint attack as a system that does not make use of a public/private key. The use of public/private keys does not directly address the issue of overprint forgeries. Additional security can be obtained by recording multiple finger prints from different locations of the original. However this approach can be overcome by the forger by extending the region of the overprint forgery to cover the entire medium or 25 all potential areas where fingerprints may have been recorded. Another method involves recording multiple fingerprints from different orientations, using a specific hash function to generate a key using the paper fingerprint of the original. Such extra security layers disadvantageously increase the computational and memory requirements of the authentication system. 30 Thus, a need exists for a system that can overcome the shortcomings of existing authentication techniques and systems, which cannot suitably detect forgeries, so that 920310 (2384075_1) -3 forgeries of documents containing authentication information can be detected. As used herein, forgery detection is a different problem to authentication detection, which cannot identify a forgery. 5 SUMMARY In accordance with an aspect of the invention, there is provided a method of detecting a forged document. A first fingerprint signature for a region of a surface of a physical medium of a document is generated using a scanner. The scanner radiates incoherent light incident to the surface of the document in a first direction to generate the 10 first fingerprint signature. A second fingerprint signature of the region of the surface of the document is generated using the scanner. The scanner radiates incoherent light incident to the surface of the document in a second direction to generate the second fingerprint signature. The second direction is different to the first direction. Information indicative of the similarity of the first and second fingerprint signatures is generated. 15 Whether the document is a forgery or not is declared dependent upon at least one similarity threshold value and the generated similarity information of the first and second fingerprint signatures. The method may further comprise the step of generating information indicative of the similarity of the first fingerprint signature and a third fingerprint signature of the 20 region. The first and third fingerprint signatures are separately generated fingerprint signatures of the region. The scanner radiates incoherent light incident to the surface of the document in the same direction to generate the third fingerprint signature as to generate the first fingerprint signature. Whether the document is declared a forgery or not is further dependent upon another similarity threshold value and the generated similarity 25 information of the first and third fingerprint signatures. The method may further comprise the step of accessing the third fingerprint signature stored in a storage device. The method may further comprise the steps of: determining a document authentication value that is a function of the generated similarity information of the first 30 and second fingerprint signatures and the generated similarity information of the first and 920310 (2384075_1) -4 third fingerprint signatures; and comparing the determined document authentication value to a document authentication threshold. The function may be subtraction. The method may further comprise the step of transforming the second fingerprint 5 signature to be in the same direction relative to the document as the first fingerprint signature, the similarity information of the first and second fingerprint signatures being generated using the first fingerprint signature and the transformed second fingerprint signature. The first fingerprint signature may be generated by imaging using the scanner the 10 region of the document to capture a first image, the first fingerprint signature being dependent upon the first image. The second fingerprint signature may be generated by imaging using the scanner the region of the document to capture a second image, the second fingerprint signature being dependent upon the second image. The step of generating similarity information of the first and second fingerprint 15 signatures may comprise comparing the captured first and second images to generate a visual similarity value. The second direction being different to the first direction may be implemented by positioning the document on the scanner at a different orientation in the second direction relative to the first direction. The first direction may be at an angle of zero (0) degrees 20 between the orientation of the document and the orientation of the scanner, and the second direction may be at an angle of 180 degrees between the orientation of the document and the orientation of the scanner. The second direction being different to the first direction may be implemented by positioning a source of the incoherent light of the scanner at a different orientation in the 25 second direction relative to the first direction. The first direction may be at an angle of zero (0) degrees between the position of the light source and the document, and the second direction may be at an angle of 180 degrees between the position of the light source and the document. The information indicative of the similarity of the first and second fingerprint 30 signatures may be generated by correlating the first and second fingerprint signatures. 920310 (2384075_1) -5 The information indicative of the similarity of the first and third fingerprint signatures may be generated by correlating the first and third fingerprint signatures. The physical medium may be paper and each fingerprint signature may be a paper fingerprint signature. 5 Generated similarity information may comprise a visual similarity value. A document declared to be a forgery may comprise a document forged by over printing a fingerprint signature on the physical medium to simulate the three-dimensional structure of an authentic physical medium on the physical medium of the forged document. 10 At least one similarity threshold value may be predetermined and dependent upon calibration information. At least two similarity threshold values may be used for declaring whether the document is a forgery or not. In accordance with another aspect of the invention, there is provided a system for 15 detecting a forged document. The system comprises a scanner, a memory for storing data and a computer program, and a processing unit coupled to the memory to execute the computer program. The processing unit and the memory are configured to detect a forged document. The computer program comprises: a computer program code module for obtaining a first fingerprint signature for a region of a surface of a physical medium of a 20 document generated by the scanner, the scanner radiating incoherent light incident to the surface of the document in a first direction to generate the first fingerprint signature; a computer program code module for obtaining a second fingerprint signature of the region of the surface of the document generated by the scanner, the scanner radiating incoherent light incident to the surface of the document in a second direction to generate the second 25 fingerprint signature, the second direction being different to the first direction; a computer program code module for generating information indicative of the similarity of the first and second fingerprint signatures; and a computer program code module for declaring whether the document is a forgery or not dependent upon at least one similarity threshold value and the generated similarity information of the first and second fingerprint 30 signatures. 920310 (2384075_1) -6 In accordance with still another aspect of the invention, there is provided a computer program product comprising a computer readable medium having recorded therein a computer program for execution by a processing unit to detect a forged document. The computer program comprises: a computer program code module for 5 obtaining a first fingerprint signature for a region of a surface of a physical medium of a document generated by the scanner, the scanner radiating incoherent light incident to the surface of the document in a first direction to generate the first fingerprint signature; a computer program code module for obtaining a second fingerprint signature of the region of the surface of the document generated by the scanner, the scanner radiating incoherent 10 light incident to the surface of the document in a second direction to generate the second fingerprint signature, the second direction being different to the first direction; a computer program code module for generating information indicative of the similarity of the first and second fingerprint signatures; and a computer program code module for declaring whether the document is a forgery or not dependent upon at least one similarity threshold 15 value and the generated similarity information of the first and second fingerprint signatures. The aspects of the invention are directed to detecting over-print forgeries. The detection is achieved using a correlation between two scans taken of a sub-region of a paper from different orientations. The correlation value is compared against a pre 20 determined threshold value to identify if the paper is a forgery or not. These and other steps may be performed to detect over-print forgeries. BRIEF DESCRIPTION OF THE DRAWINGS Embodiments of the invention are described herein with reference to the 25 drawings, in which: Figs. 1A, lB, and IC are block diagrams of a paper-fingerprint-based authentication system showing inputs and outputs of the system and the fibre structure of a paper; Figs. 2A, 2B, 2C and 2D are schematic flow and block diagrams of encoding and 30 decoding processes of a paper-fingerprint-based authentication system in accordance with an embodiment of the invention; 920310 (2384075_1) -7 Figs. 3A, 3B, 3C, 3D and 3E are schematic flow and block diagrams of the over print forgery detection process; Fig. 4 is a schematic flow diagram of the paper-fingerprint-based authentication system with forgery prevention; 5 Fig. 5 is a plot showing results of an experiment performed to determine the threshold value to classify an input sample as an authentic sample or a forgery sample; Fig. 6 is a schematic flow diagram illustrating the setup of the authentication system in accordance with a further embodiment of the invention; Fig. 7 is a block diagram illustrating embodiments of the invention; and 10 Figs. 8A and 8B are schematic block diagrams of a general-purpose computer system with which the embodiments of the invention can be practised. DETAILED DESCRIPTION 15 Methods, apparatuses, systems, and computer program products for detecting a forged document are disclosed. In the following description, numerous specific details, including particular media, light sources, alignment marks and the like are set forth. However, from this disclosure, it will be apparent to those skilled in the art that modifications and/or substitutions may be made without departing from the scope and 20 spirit of the invention. In other circumstances, specific details may be omitted so as not to obscure the invention. Where reference is made in any one or more of the accompanying drawings to steps and/or features, which have the same reference numerals, those steps and/or features have for the purposes of this description the same function(s) or operation(s), unless the 25 contrary intention appears. I. PFP Authentication A Paper Finger Print (PFP) based authentication system is described hereinafter briefly. Fig. I depicts encoding and decoding sub-systems of a PFP-based authentication 30 system. Fig. IA shows a document 101 that is sought to be protected and a rectangular region 102 of the document 101 that is pre-selected to record a PFP. The PFP may be a 920310 (2384075_1) -8 number or a set of numbers derived from the intrinsic randomness of the pre-selected region 102 of the physical medium of the document 101. The PFP is dependent on the intrinsic randomness of the structure or physical properties of the medium and a change in the intrinsic randomness greater than a certain degree results in a change in the PFP. For 5 example, if a paper medium is considered, two different paper samples generally differ in their intrinsic randomness due to different arrangement of fibres within paper samples. Hence, these paper samples have different PFPs. In the case where the PFP is simply an image of the paper medium 101 in the region 102, two PFPs might be declared to be identical if the cross-correlation between 10 the two PFPs exceeds a threshold. Otherwise, the PFPs are declared to be different. Since the particular instance of the intrinsic randomness of the medium represents a property unique to the medium, the PFP derived from that particular instance of the intrinsic randomness of the medium is also unique to the medium 101. The PFP may, for example, be simply an image of the medium that displays aspects of the intrinsic randomness of the 15 structure or physical properties of the medium. Alternatively, the PFP may be a number or a set of numbers derived from the image by a mathematical operation. One size of the rectangular region 102 that can be practiced is 1.5 inches wide by 1.5 inches long. However, any other shape and size for the region 102 can be used as well, and PFP authentication is not restricted to rectangular shapes of size 1.5 by 1.5 20 inches. The selected region is termed the PFP region and signifies the region of the document 101 used for recording the PFP. In an embodiment of the invention, the imaging technique used to collect the PFP can be implemented using a conventional document scanner set at a resolution of at least 200 dpi (preferably 600 dpi) using incoherent light for scanning. 25 In Fig. IB, the block 112 (dashed lines) represents a process of scanning a paper 109 using a conventional document scanner 110. The output is a scanned digital image 111. Other imaging techniques may be employed, however, such as a digital camera, etc. Some scanners and other imaging techniques are designed to image the textual and graphical content of a document and are not designed to image the paper underneath the 30 textual and graphical content. These imaging techniques often deliberately set the value of all pixels associated with the underlying paper to some maximum intensity value (e.g. 920310 (2384075_1) -9 255) in the captured image to maximise the number of intensity levels allocated to textual and graphical content. Such imaging techniques without modification may not be suitable for implementing the embodiments of the invention, because the unique structure of the underlying paper would be invisible. For present purposes, the imaging technique used 5 must be capable of detecting random structure on the surface and/or within the underlying medium (e.g., paper of the document). For example, an imaging system that collects an 8-bit, greyscale image of the document typically requires that the underlying paper 109 of the document 101 shows at least 2-3 bits of intensity variation. The scanned image 111 of the region 102 is used as a unique PFP for the paper 101. 10 In Fig. IC, the authenticating process based on the PFP is indicated by box 106 (dashed lines). The authentication process is divided in to two parts - encoding 107 and decoding 108. The original paper is encoded 107 to provide the PFP. A PFP and a test paper to be authenticated are provided as input to the decoding process 108 to output the authentication results. 15 II Overview of Forgery Detection Fig. 7 provides a high-level overview of techniques to detect forgeries of documents containing authentication information in accordance with two embodiments of the invention. A first process of forgery detection 700 is described, before describing 20 another forgery detection process comprising additional steps including those of process 700. Unlike authentication systems, the embodiments of the invention are able to detect forgeries that cannot be detected using authentication systems. The embodiments of the invention utilise the fact that illuminating and capturing fingerprint signatures of a physical medium from different angles will produce fingerprints with low to medium 25 correlations between them. In the embodiments of the invention, incoherent light can be practiced to implement the forgery detection. The fingerprints are dependent upon the 3D structure of the physical medium. The embodiments of the invention are focused on implementation in the decoding process 108 of Fig. 1. Paper fingerprint signatures are not required to be recorded, and only a small amount of calibration information 30 concerning a scanner and printing device may be required. In fact, the noted calibration information is not required in some embodiments. In forgeries, because the fingerprint is 920310 (2384075_1) -10 printed on the physical medium, there may be little or no 3D structure, but instead a 2D representation of the original medium masking the structure of the physical medium of the forgery. Fig. 7 depicts a method 700 of detecting a forged document in accordance with an 5 embodiment of the invention. A first fingerprint signature for a region of a surface of a physical medium of a document 710 is obtained. The first fingerprint signature 710 can be generated using a scanner (not shown) that radiates incoherent light incident to the surface of the document in a first direction to generate the first fingerprint signature. In this example, for the purposes of discussion, the paper and the scanned image are oriented 10 in the first direction at an angle of zero degrees (00 scan). A second fingerprint signature 712 of the region of the surface of the document is obtained. Again, the scanner radiates incoherent light incident to the surface of the document in a second direction to generate the second fingerprint signature. The second direction is different to the first direction. In this example, the paper and the scanned image are oriented in the second direction at 15 an angle of 180 degrees (1800 scan). A transformation 720 can be applied to the second fingerprint signature to electronically rotate the scan 712 by 180 degrees to obtain the electronically rotated scan 714. The second fingerprint signature can be transformed to be in the same direction relative to the document as the first fingerprint signature. In this embodiment, information indicative of the similarity of the first and second 20 fingerprint signatures is generated by block 730. The similarity information of the first and second fingerprint signatures is generated using the first fingerprint signature 710 and the transformed second fingerprint signature 714. The match strength between the first fingerprint signature 710 and the rotated second fingerprint signature 714 is computed by module 730, which produces as output of the block 700 a match strength B. If the match 25 strength B is low to medium, then the output indicates an authentic paper fingerprint (PFP). However, if the match strength B is medium to high, the output indicates a forged document. This embodiment is able to detect subsets of documents that are authentic or forged, but may not detect all documents due to those that have match strength B that is in their range and hence overlap. This embodiment 700 may be useful in detecting some but 30 not all documents as having authentic PFPs or forged PFPs. The output of the block 700 may be used to declare whether the document is a forgery or not dependent upon at least 920310 (2384075_1) -11 one similarity threshold value and the generated similarity information of the first and second fingerprint signatures. Fig. 7 also illustrates another method of detecting a forged document comprising blocks 700 and 750 in accordance with a further embodiment of the invention. In this 5 embodiment, an original PFP 760 is provided as input to a module 770 to compute match strength A. This original PFP 760 can be stored in a storage device (not shown) and accessed from the storage device. The module 770 generates information indicative of the similarity of the first fingerprint signature 710 and the third fingerprint signature 760 of the region. The first and third fingerprint signatures 710, 760 are separately generated 10 fingerprint signatures of the region. The match strength A output by module 770 indicates the similarity of the first and third fingerprint signatures 710 and 760. A match strength A of medium to high indicates an authentic PFP, while a match strength A of low to medium indicates a forged PFP. Thus, the computed match strength A can indicate or declare whether the document is a forgery or not. 15 A document authentication value can be determined that is a function of the generated similarity information (match strength B) of the first and second fingerprint signatures 710, 714 and the generated similarity information (match strength A) of the first and third fingerprint signatures 710, 760. Preferably, the function is subtraction. The match strength A output by module 770 can be a positive input to mixer 780 and the 20 match strength B output by module 730 is a negative input of the mixer 780. The mixer 780 provides as output the difference, A-B. This output A-B is able to better differentiate authentic and forged PFPs. The output A-B of medium indicates an authentic PFP, and a low (or negative) output A-B value indicates a forged PFP. Not shown in Fig. 7 but described hereinafter, the determined document authentication value (A-B) can be 25 compared to a document authentication threshold. The first and second fingerprint signatures 710, 712 can be generated by imaging using the scanner the region of the document to capture a first image. The first and second fingerprint signatures can be dependent upon the first and second image. The generated similarity information of the first and second fingerprint signatures comprises a 30 visual similarity value that is generated by comparing the captured first and second images. 920310 (2384075_1) -12 The second direction that is different to the first direction can be implemented by positioning the document on the scanner at a different orientation in the second direction relative to the first direction. More specifically, the first direction may be at an angle of zero (0) degrees between the orientation of the document and the orientation of the 5 scanner, and the second direction may be at an angle of 180 degrees between the orientation of the document and the orientation of the scanner. Alternatively, the second direction that is different to the first direction can be implemented by positioning a source of the incoherent light of the scanner at a different orientation in the second direction relative to the first direction. More specifically, the first direction may be at an angle of 10 zero (0) degrees between the position of the light source and the document, and the second direction may be at an angle of 180 degrees between the position of the light source and the document. The information indicative of the similarity of the first and second fingerprint signatures may be generated by correlating the first and second fingerprint signatures. 15 Likewise, the information indicative of the similarity of the first and third fingerprint signatures may be generated by correlating the first and third fingerprint signatures. The physical medium may be paper and each fingerprint signature may be a paper fingerprint signature. A document declared to be a forgery comprises a document forged by over 20 printing a fingerprint signature on the physical medium to simulate the three-dimensional structure of an authentic physical medium on the physical medium of the forged document. At least one similarity threshold value may be predetermined and dependent upon calibration information. However, at least two similarity threshold values may be used 25 for declaring whether the document is a forgery or not. These and other aspects are described in greater detail hereinafter. III. Hardware and Software Implementation Figs. 8A and 8B collectively form a schematic block diagram of a general 30 purpose computer system 800, with which the embodiments of the invention can be practiced. 920310 (2384075_1) -13 As depicted in Fig. 8A, the computer system 800 comprises a computer module 801, input devices, and output devices. Examples of input devices include a scanner 826, a camera 827 (e.g., a digital camera), a keyboard 802, a mouse pointer device 803, and a microphone 880. Further, examples of output devices include a printer 815, a display 5 device 814 and loudspeakers 817. An external Modulator-Demodulator (Modem) transceiver device 816 may be used by the computer module 801 for communicating to and from a communications network 820 via a connection 821. The network 820 may be a wide-area network (WAN), such as the Internet or a private WAN. Where the connection 821 is a telephone line, the modem 816 may be a traditional "dial-up" modem. 10 Alternatively, where the connection 821 is a high capacity (e.g., cable) connection, the modem 816 may be a broadband modem. A wireless modem may also be used for wireless connection to the network 820. The computer module 801 typically includes at least one processor unit 805 and a memory unit 806 (for example, semiconductor random access memory (RAM) and 15 semiconductor read only memory (ROM)). The module 801 also includes a number of input/output (I/O) interfaces 807, 813, 808. The audio-video interface 807 couples the module 801 with the video display 814, loudspeakers 817 and microphone 880. The I/O interface 813 couples the module 801 with the keyboard 802, the mouse 803, the scanner 826, the camera 827 and optionally a joystick (not illustrated). The interface 808 couples 20 the module 801 with the external modem 816 and the printer 815. In some implementations, the modem 816 may be incorporated within the computer module 801 (for example, within the interface 808). The computer module 801 also may have a local network interface 811 that, via a connection 823, permits coupling of the computer system 800 to a Local Area Network (LAN) 822. The local network 822 may also couple 25 to the wide network 820 via a connection 824, which would typically include a so-called "firewall" device or a device of similar functionality. The network interface 811 may be comprise an EthernetTM circuit card, a BluetoothTM wireless arrangement, an IEEE 802.11 wireless arrangement, or any of a number of known interfaces capable of connecting a device to a communications network (e.g., USB cable coupled to a cellular phone 30 functioning as a modem). 920310 (2384075_1) -14 The interfaces 808 and 813 may provide serial connectivity, parallel connectivity, or both. Serial connectivity is typically implemented according to the Universal Serial Bus (USB) standards and having corresponding USB connectors (not illustrated). Storage devices 809 are provided and typically include a hard disk drive (HDD) 810. Other 5 storage devices such as a floppy disk drive and a magnetic tape drive (not illustrated) may also be used. An optical disk drive 812 is typically provided to act as a non-volatile source of data. Portable memory devices, such optical disks (eg: CD-ROM, DVD), USB RAM, and floppy disks for example may then be used as appropriate sources of data to the system 800. 10 The components 805 to 813 of the computer module 801 typically communicate via an interconnected bus 804 and in a manner which results in a conventional mode of operation of the computer system 800 known to those in the relevant art. Examples of computers on which the described arrangements can be practised include IBM-PC's and compatibles, Sun Sparcstations, Apple MacTM or alike computer systems evolved 15 therefrom. The method of detecting a forged document may be implemented using the computer system 800 wherein the processes of Figs. I to 7 may be implemented as one or more software application programs 833 executable within the computer system 800. In particular, the steps of the method may be effected by instructions 831 in the software 833 20 that are carried out within the computer system 800. The software instructions 831 may be formed as one or more code modules, each for performing one or more particular tasks. The software may also be divided into two separate parts, in which a first part and the corresponding code modules performs the method of detecting a forged document and a second part and the corresponding code modules manage a user interface between the 25 first part and the user. The software 833 is generally loaded into the computer system 800 from a computer readable medium and is typically stored in the HDD 810 or the memory 806. After storing, the software 833 can be executed by the computer system 800. In some instances, the application programs 833 may be supplied to the user encoded on one or 30 more CD-ROM 825 and read via the corresponding drive 812 prior to storage in the memory 810 or 806. Alternatively the software 833 may be read by the computer system 920310 (2384075_1) -15 800 from the networks 820 or 822 or loaded into the computer system 800 from other computer readable media. Computer readable storage media refers to any storage medium that participates in providing instructions and/or data to the computer system 800 for execution and/or processing. Examples of such storage media include floppy disks, 5 magnetic tape, CD-ROM, a hard disk drive, a ROM or integrated circuit, USB memory, a magneto-optical disk, or a computer readable card such as a PCMCIA card and the like, whether or not such devices are internal or external of the computer module 801. Examples of computer readable transmission media that may also participate in the provision of software, application programs, instructions and/or data to the computer 10 module 801 include radio or infra-red transmission channels as well as a network connection to another computer or networked device, and the Internet or Intranets including e-mail transmissions and information recorded on Websites and the like. The second part of the application programs 833 and the corresponding code modules mentioned above may be executed to implement one or more graphical user 15 interfaces (GUIs) to be rendered or otherwise represented upon the display 814. Through manipulation of typically the keyboard 802 and the mouse 803, a user of the computer system 800 and the application may manipulate the interface in a functionally adaptable manner to provide controlling commands and/or input to the applications associated with the GUI(s). Other forms of functionally adaptable user interfaces may also be 20 implemented, such as an audio interface utilizing speech prompts output via the loudspeakers 817 and user voice commands input via the microphone 880. Fig. 8B is a detailed schematic block diagram of the processor 805 and a "memory" 834. The memory 834 represents a logical aggregation of all the memory devices (including the HDD 810 and semiconductor memory 806) that can be accessed by 25 the computer module 801 in Fig. 8A. When the computer module 801 is initially powered up, a power-on self-test (POST) program 850 executes. The POST program 850 is typically stored in a ROM 849 of the semiconductor memory 806. A program permanently stored in a hardware device such as the ROM 849 is sometimes referred to as firmware. The POST program 850 30 examines hardware within the computer module 801 to ensure proper functioning, and typically checks the processor 805, the memory 809, 806, and a basic input-output 920310 (2384075_1) -16 systems software (BIOS) module 85 1, also typically stored in the ROM 849, for correct operation. Once the POST program 850 has run successfully, the BIOS 851 activates the hard disk drive 810. Activation of the hard disk drive 810 causes a bootstrap loader program 852 that is resident on the hard disk drive 810 to execute via the processor 805. 5 This loads an operating system 853 into the RAM memory 806 upon which the operating system 853 commences operation. The operating system 853 is a system level application, executable by the processor 805, to fulfil various high level functions, including processor management, memory management, device management, storage management, software application interface, and generic user interface. 10 The operating system 853 manages the memory 809, 806 to ensure that each process or application running on the computer module 801 has sufficient memory in which to execute without colliding with memory allocated to another process. Furthermore, the different types of memory available in the system 800 must be used properly so that each process can run effectively. Accordingly, the aggregated memory 15 834 is not intended to illustrate how particular segments of memory are allocated (unless otherwise stated), but rather to provide a general view of the memory accessible by the computer system 800 and how such is used. The processor 805 includes a number of functional modules including a control unit 839, an arithmetic logic unit (ALU) 840, and a local or internal memory 848, 20 sometimes called a cache memory. The cache memory 848 typically includes a number of storage registers 844 - 846 in a register section. One or more internal buses 841 functionally interconnect these functional modules. The processor 805 typically also has one or more interfaces 842 for communicating with external devices via the system bus 804, using a connection 818. 25 The application program 833 includes a sequence of instructions 831 that may include conditional branch and loop instructions. The program 833 may also include data 832 which is used in execution of the program 833. The instructions 831 and the data 832 are stored in memory locations 828-830 and 835-838 respectively. Depending upon the relative size of the instructions 831 and the memory locations 828-830, a particular 30 instruction may be stored in a single memory location as depicted by the instruction shown in the memory location 830. Alternately, an instruction may be segmented into a 920310 (2384075_1) -17 number of parts each of which is stored in a separate memory location, as depicted by the instruction segments shown in the memory locations 828-829. In general, the processor 805 is given a set of instructions which are executed therein. The processor 805 then waits for a subsequent input, to which it reacts to by 5 executing another set of instructions. Each input may be provided from one or more of a number of sources, including data generated by one or more of the input devices 802, 803, data received from an external source across one of the networks 820, 822, data retrieved from one of the storage devices 806, 809 or data retrieved from a storage medium 825 inserted into the corresponding reader 812. The execution of a set of the 10 instructions may in some cases result in output of data. Execution may also involve storing data or variables to the memory 834. The software implementing the method of detecting a forged document uses input variables 854, which are stored in the memory 834 in corresponding memory locations 855-858. The software also has output variables 861, which are stored in the memory 15 834 in corresponding memory locations 862-865. Intermediate variables may be stored in memory locations 859, 860, 866 and 867. The register section 844-846, the arithmetic logic unit (ALU) 840, and the control unit 839 of the processor 805 work together to perform sequences of micro-operations needed to perform "fetch, decode, and execute" cycles for every instruction in the 20 instruction set making up the program 833. Each fetch, decode, and execute cycle comprises: (a) a fetch operation, which fetches or reads an instruction 831 from a memory location 828; (b) a decode operation in which the control unit 839 determines which instruction 25 has been fetched; and (c) an execute operation in which the control unit 839 and/or the ALU 840 execute the instruction. Thereafter, a further fetch, decode, and execute cycle for the next instruction may be executed. Similarly, a store cycle may be performed by which the control unit 839 30 stores or writes a value to a memory location 832. 920310 (2384075_1) -18 Each step or sub-process in the processes of Figs. I to 7 is associated with one or more segments of the program 833, and is performed by the register section 844-847, the ALU 840, and the control unit 839 in the processor 805 working together to perform the fetch, decode, and execute cycles for every instruction in the instruction set for the noted 5 segments of the program 833. The method of detecting a forged document may alternatively be implemented in dedicated hardware such as one or more integrated circuits performing the functions or sub functions of detecting a forged document. Such dedicated hardware may include graphic processors, digital signal processors, or one or more microprocessors and 10 associated memories. Figs. 2A, 2B, and 2D are flow-charts of the encoding and decoding processes (107, 108 of Fig. IC). Fig. 2C is a block diagram illustrating scanned images resulting from scanning of a test paper at different orientations. In Fig. 2A, the PFP encoding process 200 takes a sheet of paper as input and 15 generates a PFP. Processing commences in step 201. In step 201, the paper is scanned using a standard scanner (826 in Fig. 8A). In step 202, the PFP is generated from the scanned image of the paper. The scanned image may be stored in memory 806 in Fig. 8 via I/O interface 813 and bus 804. The processor can retrieve the scanned image from the memory 806 and generate the PFP, which can then be stored in the memory 806, for 20 example. Processing then terminates. Referring to Fig. 2B, the method 203 explains in greater detail the process carried out in step 202 of Fig. 2A, i.e. generation of the PFP. The method 203 of generating a PFP from the scanned image commences in step 204. In step 204, a coordinate-system is established to select the pre-defined area (as represented by 102 in Fig. IA). The process 25 of step 204 is also known as registration and may be performed using a Rotation-Scale Translation alignment using pre-defined alignment marks. Step 204 may be implemented using the processor 805. The alignment marks may be the page boundaries, e.g. the top left point of the page acts as the origin of the coordinate system. Alternatively, the alignment marks may be any pre-existing marks on the page used as anchor points to 30 define a coordinate system and align the scanned image with respect to the coordinate system. In step 205, a PFP is recorded. More specifically, to record a PFP, a predefined 920310 (2384075_1) -19 region is selected with respect to the coordinate system and the scanned image values in the predefined region are output by the processor 805 as the PFP for the input document. Processing then terminates. In embodiments of the invention, the PFP for a (test) paper 210 in Fig. 2C is 5 recorded at an orientation angle of 0 degrees, where 0 degrees refers to a default orientation. For example, the paper is placed at zero degree orientation in the scanner relative to the orientation of the scan image. The scan image 211 of the test paper 210 is depicted in Fig. 2C, where both are oriented the same for the zero-degree orientation. Similarly, the scan image 212 shows an example of a scan image if where the (test) paper 10 210 is scanned at an orientation of 180 degree relative to the default zero-degree orientation. Fig. 2D illustrates in detail the PFP decoding method 220 in accordance with an embodiment of the invention. Processing commences at step 230. Steps 230 and 232 in Fig. 2D are similar to steps 201 and 202 in the encoding process 200 shown in Fig 2A. In 15 step 230, the document to test 280 is input, and the paper is scanned. Again, this may be done by the scanner 826 in Fig. 8A. In step 232, the PFP is generated by the processor 805 from the scanned image. In step 234, the generated PFP from step 232 and the reference PFP 282 input to step 234 are correlated. Step 234 uses the processor 805 to compare the generated PFP with the reference PFP. The reference PFP is generated at 20 some previous time using the encoding method 200 of Fig. 2A and may be stored in memory 806. The reference PFP 282 is used as a baseline reference, i.e. to check if the generated PFP matches with the reference PFP 282 or not. In general, the reference PFP 282 is generated by the method 200 at some time in the past when the original document 25 was created. The reference PFP 282 is assumed to be the valid PFP for the authentic original medium. Subsequently generated PFPs can be compared to the reference PFP 282 to authenticate the medium. After creation, the reference PFP is stored, e.g. either on the medium itself or external to the medium in, for example, a computer database in memory 806 or the HDD. The reference PFP 282 is stored in such a way as to allow step 30 234 to access the reference PFP 282. Step 234 uses a correlation process implemented by 920310 (2384075_1) -20 processor 805, for example cross-correlation, to compare the generated PFP with the reference PFP 282. In step 234, cross-correlation is not the only method that can practiced to compare two (or more) signatures to each other (in Fig. 2D, the verb "compare" is listed 5 parenthetically to indicate this). Various mathematical equations can be used. For example, the dot-product, or the Euclidean norm, between the images (signatures) can be computed and is well known to those skilled in the art. In addition, comparison methods can compare images while ignoring certain aspects of the images. For example, two images (signatures) with the textual information, graphical information, image 10 information, machine readable code, and/or other forms of content printed onto the media may be compared while ignoring the printed information in the comparison operation. If the signatures are stored in a binary representation, the Hamming distance can be used to perform the comparison. In general, the exact comparison function depends on the mathematical formulae used to create the signature. 15 In decision step 236, the correlation output from step 234 is compared using the processor 805 with a pre-defined threshold value. If the correlation value is greater than the threshold in step 236 (Yes), the result "Authentic" is provided or output in step 240. That is, the test paper 280 is declared to be an authentic sample with respect to the reference PFP 282. Otherwise, if step 236 returns false (No), the result "Not Authentic" 20 is provided or output in step 238. That is, the test paper sample 280 is declared not to match with the reference PFP 282. From steps 238 and 240, processing terminates. V. Scanning two scans and using match strength to identify forgery An example of an over-print method is to print a pattern on the paper using a 25 printer in a pre-defined region. In particular, an image of a PFP is printed onto the PFP zone on the forgery. When the forgery is scanned, the printed image of the PFP correlates with the reference PFP, potentially giving the false impression of a match of the underlying paper fibre structure with the reference PFP. The over-print forgery method can defeat existing PFP systems, because the suitably high correlation value can be 30 obtained. This is mainly due to advancements in printing technology that allow printing a reference PFP on to a test paper such that the PFP generated from the test paper gives the 920310 (2384075_1) -21 high enough correlation value to be considered as an authentic sample. In fact the correlation is primarily due to the printed image, and not the paper fibre structure of the forgery. The embodiments of the invention aim to detect if an over-print method has been 5 used to create a forgery. Fig. 3A illustrates at a high-level a method 300 of detecting an over-print forgery. Processing commences in step 320. Step 320 represents a process that generates a correlation value for an input document 310, which is being tested for forgery detection. This step may be implemented using processor 805. Step 320 is described in detail 10 hereinafter with reference to Fig. 3B. In decision step 322, the correlation value from step 320 for the document being tested 3 10 is compared using the processor 805 with a pre-defined threshold. A threshold value = 20.0 may used. However, other threshold values can be practiced. One way to determine the threshold value is by analysing many examples of over-print forgeries and original samples and choosing a threshold such that 15 the number of false matches is minimized while the number of false rejections is also kept low, or as low as possible. If the correlation value is higher than the threshold (YES), processing continues at step 326, where it is declared to be a forgery ("Over-print Present") using the processor 805. Otherwise, if decision step 322 returns false (NO), i.e. the correlation value is less than or equal to the threshold, processing continues at step 20 324, where is declared not to be a forgery ("No Over-print") using the processor 805. From steps 324 and 325, processing terminates. In some applications, more than one threshold may be applied for discriminating between an authentic PFP and a forged PFP. Preferably in some applications two threshold levels may be used. The first threshold level is the higher threshold and the 25 second threshold level is the lower threshold. Fig. 3E illustrates a process 330 of using two threshold levels. Processing commences in decision step 341 which compares the correlation value with the first threshold value. If the correlation value is higher than the first threshold level (YES) processing continues at step 342, where it is declared to be a forgery ("Over-print Present) using the processor 805. Processing then terminates. 30 Otherwise, if decision step 341 returns false (NO) i.e. the correlation value is less than or equal to the first threshold value, processing continues at decision step 343. In decision 920310 (2384075_1) -22 step 343, the correlation value is compared with the second threshold value. If the correlation value is lower than the second threshold (YES), processing continues at step 344 where it is declared not to be a forgery ("No Over-print"). Processing then terminates. Otherwise, if decision step 343 returns false (NO), processing continues at 5 step 345 where it is declared that no decision is made whether over-print forgery is present or not. Processing then terminates. Step 320 of Fig. 3A is depicted in greater detail in Fig. 3B. Processing commences at step 301 in Fig. 3B. In step 301, a first scan is taken, or collected, of an input paper in a first direction, as depicted with reference to Fig. 3C. Again, this may be 10 done using scanner 826 in Fig. 8A. The first direction is selected to be zero (0) degrees preferably, where 0 degrees is consistent with the default orientation established in Fig 2C. Referring again to Fig. 3C, the physical test paper 308 is input. Crosses 309 on the paper 308 represent alignment marks printed on the paper 308 for registration. Scanning at angle 0 in block 311 of Fig. 3C means that the test paper is placed on the scanner bed 15 so that the scan image output is similar to the image 314. The term 'scan at angle 0' does not mean to put the test paper on the scanner bed precisely at 0 degrees, but instead the placement should be approximate within usual human visual precision. A very fine alignment is not required as the PFP generation process involves a registration step as discussed herein before. The test paper need not be placed at angle of zero degrees. 20 Different angles may be practiced. In step 302, a first PFP is generated from the first scan. Step 302 is similar to step 202 of Fig. 2A which registers the scanned image using the alignment marks and generates the PFP. The PFP obtained from the scan taken at angle 0 is termed as PFP_0. In step 303 of Fig. 3, a second scan is taken, or collected, as shown at a different 25 orientation in Fig. 3D. A 180-degree orientation may be used. The test paper 308 is scanned at an angle of 180 degrees in block 312 of Fig. 3D to output an inverted scan image 313 of the test paper 308. However, other orientations such as 270 degrees may be practiced, and the embodiments of the invention are not restricted to 180 degree orientation. In Fig. 3D, the test paper 308 is placed on the scanner bed in a rotated 30 fashion such that the output scan image 313 as generated by the scanner is rotated by 180 degrees relative to the scan used to obtain PFP_0. 920310 (2384075_1) -23 Referring again to Fig. 3B, processing continues at step 304, which is also similar to step 232 that registers the scanned image and generates the PFP. The PFP obtained from the 180-degree oriented scan is termed as PFP_180. In step 305, a correlation value is generated by correlating PFPs PFP_0 and 5 PFP_180. The correlation value is generated by the processor 805 and is the output of step 305, which is input to the decision step 322 of Fig. 3A. Processing then terminates. VI. Two scans used in addition to 3 d scan stored in database In this embodiment, the forgery detection process uses the PFP for the reference 10 document, that is, the original document for which any new documents is sought to be authenticated against. Fig 6 illustrates this embodiment. A reference document 601 is input having a PFP that is captured using a PFP-generating process 602. The process 602 refers to the paper encoding process 200 shown in Fig. 2. The PFP of the reference document is termed the referencePFP. The referencePFP is provided to the system 15 either using a database or by any other means such as from a storage device or by communication over a network etc as indicated by the block 603 in Fig. 6. The processing block 604 takes the test document 605 and the referencePFP from the storage network 603 as input to detect if the test document 605 is a forgery or not. The test document as 605 refers to a document to be tested to determine if that document is the reference 20 document or not. The forgery detection process 604 is described in greater detail with reference to Fig. 4. The process 604 is illustrated in more detail in Fig. 4. The method 400 shown in Fig. 4 is an authentication process with forgery prevention. Processing commences in steps 401 and 403. Step 401 generates a correlation value from the input document to test 25 402. The step 401 is the same as the process 320 of Fig 3A, which generates a correlation value between two PFPs obtained at 0 and 180 degrees. The output correlation value is labelled CorrB. From step 401, processing continues at step 405. As shown in Fig. 4, step 403 is performed in parallel with step 401. The document to be tested 402 is also input to step 403. Step 403 is similar to process 202 of 30 Fig. 2A, which generates a PFP from the input document by placing the document at 0 degree orientation and termed as PFP_0. 920310 (2384075_1) -24 From step 403, processing continues at step 404. In step 404, the PFP at 0 degree orientation PFP_0 is correlated with the reference_PFP, which is an input to step 404. The output correlation value is labelled CorrA. The correlation values CorrA and CorrB are input to step 405. From step 404, 5 processing also continues at decision step 406. In step 405, a value F is calculated based on a functionf from CorrA and CorrB: F =f (CorrA, CorrB). An example function which is used in this embodiment is subtraction, i.e.: F = Corr_A - CorrB. 10 However, other functions may be practised; a suitable functionf would be expected to have a monotonic relationship with subtraction. The value F produced in step 405 is an input to decision step 407 described hereinafter. From step 404, processing continues at decision step 406. Step 406 compares the value CorrA with a predefined threshold (referred as threshold_1 in Fig 4) to determine 15 if CorrA is greater than or equal to the threshold_1. In this embodiment,threshold_1 = 10.0 is used. However, other values of threshold_ I can be practiced such as 20.0.. If CorrA is less than threshold_1 (NO), processing continues at step 410 and the test paper 402 is declared to be 'not authentic'. From step 410, processing terminates. If CorrA is greater than or equal to the threshold_1 in decision step 406 (YES), 20 processing continues at decision step 407. Step 407 compares the F value generated by step 405 with another predefined forgery threshold (referred as threshold_2 in Fig 4). The forgery threshold=20.0 is used in this embodiment. However, other values of threshold_2 can be practiced. Fig. 5 shows results of an experiment performed to determine the threshold value 25 for forgery threshold (i.e. threshold_2). The X axis 501 represents the number of samples used in the experiment and goes from 1 to 100 covering 100 samples shown in this figure. The Y axis 502 plots the F value. Two types of sample conditions, authentic samples 503 and forgery samples 504, are used. Authentic samples 503 refer to samples when the test document is indeed the reference document. Forgery samples 504 refer to samples where 30 the test document is different than the reference document and an overprint forgery attempt is tried to match the test document with the reference document. As Fig. 5 shows, 920310 (2384075_1) -25 a threshold value = 20 well separates authentic samples 503 and forgery samples 504. There are other statistical techniques to identify a threshold to classify input samples, such as Naive Bayes classification, etc., which are known to those skilled in the art. Any of these classification techniques may be used to classify the input sample as a forgery or 5 an authentic sample based on its F value. Referring again to step 407 of Fig. 4, if the F value is higher or equal to the forgery threshold (NO), processing continues at step 412 and the test document 402 is declared as a forgery. Processing then terminates. Otherwise, if the F value is lower than the forgery threshold (YES), processing continues at step 414 and the test document is 10 declared to be authentic. Processing then terminates. VII. Other forgery prevention methods Other forgery prevention methods such as Linear Regression, Frequency method and combined with error correction codes etc may be practiced. A further embodiment 15 uses the forgery prevention method described above in combination with other methods to realise an authentication system with a strong resistance to forgery methods, including the over-print method of forgery creation. The encoding process generates the paper fingerprint of the document, which is stored in a storage device. At the encoder side, for a given authentication system 20 (document type, scanner type and printer type), a linear regression analysis is performed. The linear regression analysis analyses the linear relationship between PFP code values for the original PFP, after registration, and the testPFP_0. The linear regression analysis fits an equation of a line to the relationship between the two scans of the same authentic PFP code values. This relationship is computed by considering a number of pairs of 25 matching example PFPs from different scans of the same document, for one or more source documents. The documents may be on the same paper stock and scanned using the same scanner, but this need not be the case. A number of properties of the linear best fit can be used to distinguish rescan of authentic PFPs from the forgeries. In this embodiment, the following three properties are used: 920310 (2384075_1) -26 e Eigenvalue Ratio (ER): The ratio(4/ 2 ) -1 , where A and A 2 are the first and second eigenvalues, respectively. Two PFPs belonging to the same document are expected to have a high ER. " Slope: The slope of the linear fit should be close to 1.0 for two PFPs belonging to 5 the same document. " Offset (or mindist): The distance from (0, 0) of the intercept of the linear fit with the x/y axis. This value is close to 0.0 for two PFPs belonging to the same document. These values are combined to a single parameter using a Mean Property Vector and a 10 Property Covariance Matrix: e Mean Property Vector (pp): A three element vector containing the mean values of slope, offset and ER for authentic PFPs using the target paper stock. * Property Covariance Matrix (Cp): A three-by-three-element matrix, where the ijth element is the covariance between the ith and jth property of lines of best fit for 15 the target paper stock, as stored in the mean property vector. The diagonal elements of this matrix are the variances of each of the properties and the off diagonal elements quantify how one property is expected to change when another property changes. When a test document is presented for authentication, a best fit line for the 20 relationship between code values in the authentic PFP and those in the test PFP is computed. A Linear Regression (LR) distance metric is calculated using a Malhalanobis distance metric, a distance metric well known to those skilled-in-the-art, using the following expression: Di, = (x -p,) T C I(x -- p,), 25 where the elements of x contain the properties of the best fit line calculated for the test document. The value of Dir is compared with a pre-defined threshold. If Dir is higher than the threshold value, the sample is declared a forgery, but otherwise is authentic. The threshold DI, can be determined by examining the detector performance for sets of 920310 (2384075_1) -27 original and overprint forged PFPs and setting the threshold such that no overprint forged PFPs are accepted as authentic with the minimum number of original PFP being rejected. The authentication system can further be strengthened using known methods, such as: 5 e Frequency analysis to identify the class of the document type used. This detects if a different paper type is used to create a forgery. Additionally, frequency analysis also detects if the frequency response of the PFP region is changed by printing on the PFP region in an attempt to create a forgery. 0 Hash Key - A hash function may be used on the PFP to generate a key, which is 10 then encrypted using a private/public encryption. This encrypted message is stored on the document. This measure increase resistance against forgeries further by creating the need to know the hash function and the requirement of private key for encryption. Methods, apparatuses, systems, and computer program products for detecting a 15 forged document have been disclosed. The foregoing describes only some embodiments of the present invention, and modifications and/or changes can be made thereto without departing from the scope and spirit of the invention, the embodiments being illustrative and not restrictive. In the context of this specification, the word "comprising" means "including 20 principally but not necessarily solely" or "having" or "including", and not "consisting only of'. Variations of the word "comprising", such as "comprise" and "comprises" have correspondingly varied meanings. 920310 (2384075_1)

Claims (21)

1. A method of detecting a forged document, the method comprising the 5 steps of: generating using a scanner a first fingerprint signature for a region of a surface of a physical medium of a document, the scanner radiating incoherent light incident to the surface of the document in a first direction to generate the first fingerprint signature; generating using the scanner a second fingerprint signature of the region of the 10 surface of the document, the scanner radiating incoherent light incident to the surface of the document in a second direction to generate the second fingerprint signature, the second direction being different to the first direction; generating information indicative of the similarity of the first and second fingerprint signatures; and 15 declaring whether the document is a forgery or not dependent upon at least one similarity threshold value and the generated similarity information of the first and second fingerprint signatures.

2. The method as claimed in claim 1, further comprising the step of generating information indicative of the similarity of the first fingerprint signature and a 20 third fingerprint signature of the region, said first and third fingerprint signatures being separately generated fingerprint signatures of the region, the scanner radiating incoherent light incident to the surface of the document in the same direction to generate the third fingerprint signature as to generate the first fingerprint signature; and wherein whether the document is declared a forgery or not is further dependent upon another similarity 25 threshold value and the generated similarity information of the first and third fingerprint signatures.

3. The method as claimed in claim 2, further comprising the step of accessing said third fingerprint signature stored in a storage device.

4. The method as claimed in claim 2, further comprising the steps of: 920310 (2384075_1) -29 determining a document authentication value that is a function of the generated similarity information of the first and second fingerprint signatures and the generated similarity information of the first and third fingerprint signatures; and comparing the determined document authentication value to a document 5 authentication threshold.

5. The method as claimed in claim 4, wherein the function is subtraction.

6. The method as claimed in claim 1 or 2, further comprising the step of transforming the second fingerprint signature to be in the same direction relative to the document as the first fingerprint signature, the similarity information of the first and 10 second fingerprint signatures being generated using the first fingerprint signature and the transformed second fingerprint signature.

7. The method as claimed in claim 1, wherein: said first fingerprint signature is generated by imaging using the scanner the region of the document to capture a first image, said first fingerprint signature being 15 dependent upon said first image; and said second fingerprint signature is generated by imaging using the scanner the region of the document to capture a second image, said second fingerprint signature being dependent upon said second image.

8. The method as claimed in claim 7, wherein the step of generating 20 similarity information of the first and second fingerprint signatures comprises comparing the captured first and second images to generate a visual similarity value.

9. The method as claimed in claim 1, wherein the second direction being different to the first direction is implemented by positioning the document on the scanner at a different orientation in the second direction relative to the first direction. 25

10. The method as claimed in claim 9, wherein the first direction is at an angle of zero (0) degrees between the orientation of the document and the orientation of the scanner, and the second direction is at an angle of 180 degrees between the orientation of the document and the orientation of the scanner.

11. The method as claimed in claim 1, wherein the second direction being 30 different to the first direction is implemented by positioning a source of said incoherent 920310 (2384075_1) -30 light of the scanner at a different orientation in the second direction relative to the first direction.

12. The method as claimed in claim 11, wherein the first direction is at an angle of zero (0) degrees between the position of the light source and the document, and 5 the second direction is at an angle of 180 degrees between the position of the light source and the document.

13. The method as claimed in claim 1, wherein the information indicative of the similarity of the first and second fingerprint signatures is generated by correlating the first and second fingerprint signatures. 10

14. The method as claimed in claim 2, wherein the information indicative of the similarity of the first and third fingerprint signatures is generated by correlating the first and third fingerprint signatures.

15. The method as claimed in claim 1, wherein the physical medium is paper and each fingerprint signature is a paper fingerprint signature. 15

16. The method as claimed in claim 1, wherein generated similarity information comprises a visual similarity value.

17. The method as claimed in claim 1, wherein a document declared to be a forgery comprises a document forged by over-printing a fingerprint signature on said physical medium to simulate the three-dimensional structure of an authentic physical 20 medium on said physical medium of said forged document.

18. The method as claimed in claim 1, wherein at least one similarity threshold value is predetermined and dependent upon calibration information.

19. The method as claimed in claim 1, wherein at least two similarity threshold values are used for declaring whether the document is a forgery or not. 25

20. A system for detecting a forged document, the system comprising: a scanner; a memory for storing data and a computer program; and a processing unit coupled to said memory to execute said computer program, said processing unit and said memory configured to detect a forged document, said computer 30 program comprising: 920310 (2384075_1) -31 computer program code means for obtaining a first fingerprint signature for a region of a surface of a physical medium of a document generated by said scanner, the scanner radiating incoherent light incident to the surface of the document in a first direction to generate the first fingerprint signature; 5 computer program code means for obtaining a second fingerprint signature of the region of the surface of the document generated by said scanner, the scanner radiating incoherent light incident to the surface of the document in a second direction to generate the second fingerprint signature, the second direction being different to the first direction; computer program code means for generating information indicative of the 10 similarity of the first and second fingerprint signatures; and computer program code means for declaring whether the document is a forgery or not dependent upon at least one similarity threshold value and the generated similarity information of the first and second fingerprint signatures.

21. A computer program product comprising a computer readable medium 15 having recorded therein a computer program for execution by a processing unit to detect a forged document, the computer program comprising: computer program code means for obtaining a first fingerprint signature for a region of a surface of a physical medium of a document generated by a scanner, the scanner radiating incoherent light incident to the surface of the document in a first 20 direction to generate the first fingerprint signature; computer program code means for obtaining a second fingerprint signature of the region of the surface of the document generated by said scanner, the scanner radiating incoherent light incident to the surface of the document in a second direction to generate the second fingerprint signature, the second direction being different to the first direction; 25 computer program code means for generating information indicative of the similarity of the first and second fingerprint signatures; and computer program code means for declaring whether the document is a forgery or not dependent upon at least one similarity threshold value and the generated similarity information of the first and second fingerprint signatures. 920310 (2384075_1) -32 DATED this 13th Day of November 2009 CANON KABUSHIKI KAISHA Patent Attorneys for the Applicant 5 SPRUSON&FERGUSON 920310(2384075_1)