METHODS AND SYSTEMS FOR VERIFYING SIGNATURES ON SITE

Abstract

The present invention relates to a method and system for detecting counterfeit writable documents, comprising encrypting a reference handwriting sample, compressing the reference handwriting sample, embedding the reference handwriting sample in the writable document, receiving the writable document containing a handwritten segment, encrypting the handwritten segment of the received writable document, comparing the encrypted reference handwriting sample to the encrypted handwritten segment, and evaluating the probability that the encrypted reference handwriting sample and the encrypted handwritten segment were written by the same person based on the comparison. Figure for abstract: [FIG.2]

Description

Title of the invention: METHODS AND SYSTEMS FOR VERIFYING SIGNATURES ON SITE Background

[0001] Signature or handwritten sample verification is a technique used by various entities, such as banks, to verify and validate an individual's identity. Signature or handwritten sample verification can be used to compare signatures or handwritten samples with previously captured signatures or handwritten samples. An image of a signature or handwritten sample, or an actual signature or handwritten sample, can be visually inspected or submitted to verification software for comparison with the previously captured image stored in a file. Signature or handwritten sample verification software is a type of software that compares signatures or handwritten samples and determines whether the signature or handwritten sample is authentic. Summary

[0002] In one aspect, the technology relates to a method for detecting counterfeiting of a writable document comprising the encryption of a reference writing sample, the compression of the encrypted reference writing sample, the integration of the encrypted and compressed reference writing sample into the writable document, the receipt of the writable document, the writable document containing a handwritten segment, the encryption of the handwritten segment of the received writable document, the comparison of the encrypted reference writing sample with the encrypted handwritten segment, and the evaluation of a probability that the encrypted reference writing sample and the encrypted handwritten segment were written by the same person based on the comparison.

[0003] In one example, the encryption of the reference handwriting sample involves encrypting one of an image of the reference handwriting sample and a biometric feature of the reference handwriting sample. In another example, the encryption of the reference handwriting sample involves encrypting one of an image of a signature and a biometric feature of the signature. In yet another example, at least one of the encryption, compression, and comparison steps involves the use of a neural network. In some examples, the integration of the encrypted and compressed reference handwriting sample into the writable document involves generating on the writable document at least one of a machine-readable tag. or by a human operator, a printed label, a printed barcode, a printed quick response code, a magnetic stripe, and a radio-frequency identification tag. In other examples, receiving the writable document involves receiving it at one of two locations: an automated teller machine (ATM) and a receiving institution. The receiving institution may be a financial institution. In still other examples, receiving the writable document involves scanning at least the handwritten portion of the received writable document.

[0004] In further examples, the writable document includes one of a check and one of a legal instrument such as a contract, will, or other agreement. In yet another example, the encryption of the handwritten segment included in the writable document involves encrypting one of an image of the handwritten segment and one of a biometric characteristic of the handwritten segment. In yet another example, the probability assessment involves generating a similarity factor between the encrypted reference handwriting sample and the encrypted handwritten segment. In still other examples, the method further involves establishing a similarity factor threshold, the encrypted reference handwriting sample and the encrypted handwritten segment being considered to have been written by the same person when the generated similarity factor is equal to or greater than the established similarity factor threshold.

[0005] In another aspect, the technology relates to a system for detecting counterfeit writable documents which includes an updatable data repository, an integration device and a computer device operationally coupled to the updatable data repository and the integration device, the computer device having a processor and a memory, the memory storing instructions which, when executed by the processor, perform a set of operations including, via the processor, the encryption of a reference writing sample, via the processor, the compression of the encrypted reference writing sample, and via the integration device, the integration of the encrypted and compressed reference writing sample into the writable document.

[0006] In one example, the processor includes a neural network. In another example, the instruction set includes encrypting the reference writing sample by encrypting one of an image of the reference writing sample and a biometric feature of the reference writing sample. In yet another example, the instruction set includes encrypting the reference writing sample by encrypting one of an image of a signature and a biometric feature of the signature. In other examples, the set of operations includes integrating the encrypted reference writing sample and compressed into the writable document by generating at least one of a readable tag, a printed label, a printed barcode, a printed quick response code and a radio frequency identification tag on the writable document.

[0007] In yet another aspect, the technology relates to a system for detecting counterfeit writable documents comprising a receiving device, an updatable data repository functionally coupled to the receiving device, and a computer device operationally coupled to the receiving device and the updatable data repository, the computer device comprising a memory, the memory storing instructions which, when executed by the processor, perform a set of operations comprising the reception of the writable document on the receiving device, the writable document comprising a handwritten segment and an encrypted and compressed reference writing sample integrated therein, via the processor, the encryption of the handwritten segment of the writable document, via the processor,The comparison of the encrypted reference handwriting sample to the encrypted handwritten segment and, via the processor, the evaluation of the probability that the encrypted reference handwriting sample and the encrypted handwritten segment were written by the same person based on the comparison.

[0008] In one example, the processor comprises a neural network. In another example, the writable document comprises one of a check and one of a legal instrument. In yet another example, the receiving device comprises one of an automated teller machine and one of a reading machine at a receiving institution. In yet another example, the set of operations comprises receiving the writable document on the receiving device by scanning at least the handwritten segment of the writable document received on the receiving device. In yet another example, the set of operations comprises encrypting the handwritten segment included in the writable document by encrypting one of an image of the handwritten segment and one of a biometric characteristic of the handwritten segment.In other examples, the set of operations includes evaluating the probability by generating a similarity factor between the encrypted reference handwriting sample and the encrypted handwritten segment. In still other examples, the set of operations further includes establishing a similarity factor threshold, with the encrypted reference handwriting sample and the encrypted handwritten segment being considered to have been written by the same person when the generated similarity factor is equal to or greater than the established similarity factor threshold.

[0009] In another additional aspect, the technology relates to a system for detecting counterfeit writable documents which includes a receiving device, an integration device, and at least one updatable data repository functionally coupled to the receiving device and the integration device, at least one computer device operationally coupled to the receiving device and at least one updatable data repository, at least one computer device comprising at least one processor and at least one memory, the memory storing instructions which, when executed by at least one processor, perform a set of operations comprising, via at least one processor, the encryption of a reference writing sample, via at least one processor, the compression of the encrypted reference writing sample, via the integration device, the integration of the encrypted and compressed reference writing sample into the writable document, the reception of the writable document on the receiving device, the writable document comprising a handwritten segment, via at least one processor,the encryption of the handwritten segment of the writable document, via at least one processor, the comparison of the encrypted reference handwriting sample to the encrypted handwritten segment, and via at least one processor, the evaluation of the probability that the encrypted reference handwriting sample and the encrypted handwritten segment were written by the same person based on the comparison.

[0010] In one example, at least one of the reference handwriting sample and the handwritten segment contains a signature. In another example, the processor includes a neural network. In yet another example, the instruction set includes encrypting the reference handwriting sample by encrypting one of an image of a signature and a biometric feature of the signature. In yet another example, the instruction set includes encrypting the reference handwriting sample by encrypting one of an image of the reference handwriting sample and a biometric feature of the reference handwriting sample. In yet another example, the set of operations includes encrypting the handwritten segment of the writable document by encrypting one of an image of the handwritten segment and a biometric feature of the handwritten segment.

[0011] Details of one or more techniques are shown in the accompanying drawings and the description below. Other features, objects, and advantages of these techniques are clearly apparent from the description, drawings, and claims. Brief description of the drawings

[0012] Fig. 1 is a diagram illustrating a system for detecting counterfeit writable documents, in accordance with various aspects of this disclosure.

[0013] Fig. 2 is an illustration of a writable document with an embedded reference writing sample, in accordance with various aspects of this disclosure.

[0014] Fig. 3 is a diagram illustrating a system for receiving a handwriting sample or a reference signature to be authenticated, in accordance with various aspects of this disclosure.

[0015] Fig. 4 is a diagram illustrating a system for detecting counterfeit writable documents, in accordance with various aspects of this disclosure.

[0016] Fig. 5 is a flowchart illustrating a method for detecting counterfeit of a writable document, in accordance with various aspects of this disclosure.

[0017] Fig. 6 represents a schematic diagram of a computer device. Detailed description

[0018] Visually determining whether a signature or handwritten sample is authentic presents several challenges, particularly when the determination of authenticity must be performed in real time. One challenge lies in the fact that authentic signatures or handwritten samples from the same person may exhibit a degree of variability, for example, over time or depending on a given state of health. Another challenge lies in the fact that the degree of variability between authentic signatures or handwritten samples may vary from one person to another.Another challenge lies in the fact that forgers imitate signatures or handwritten samples in an attempt to obtain money, goods, and / or services. Even with sophisticated forgeries, it can be difficult to determine with the naked eye whether a signature or handwritten sample is genuine or a counterfeit. Therefore, the ability to improve the accuracy of a real-time verification process can be advantageous. Consequently, it is beneficial to reliably and accurately verify a signature or handwritten sample to determine whether the provided signature or handwritten sample is genuine or a counterfeit. Another advantage is the ability to reliably and accurately verify the signature or handwritten sample in real time, for example, when someone presents a document such as a check to a financial institution.Consequently, there is a technical problem arising from the fact that verifying signatures or handwritten samples, in real time for example, is very time-consuming and inefficient.

[0019] Signature and handwriting verification technologies have become important tools for ensuring the integrity and security of financial transactions and legal documents. Generally, archived historical data is used as a reference point for comparison. For example, an archived handwritten sample or signature can be used as a reference to compare the signature or handwritten sample on a check, contract, etc. wills, handwritten messages, and the like. However, a significant challenge arises in scenarios such as automated teller machines (ATMs) and check processing, where check forgery detection is hampered by the lack of archived authentic signature samples for signature verification, or archived genuine handwritten samples for check forgery detection. The absence of these authentic samples undermines the effectiveness of traditional signature and handwritten sample verification techniques, leaving institutions such as financial institutions or other entities requiring handwriting verification vulnerable to fraudulent activity.It is therefore advantageous, even necessary, to have a solution that allows for the provision of authentic signature samples or handwritten samples directly at the point of verification, for example, at an ATM, a notary's office, or a bank branch. In particular, encoding the authentic signatures of account holders onto the writable documents themselves, such as a check, a contract, or a will, in order to compare the encoded authentic signatures with the handwriting on the same writable document, for example, a check, can provide an advantageous method for detecting signature forgery, check forgery, or other types of counterfeiting.

[0020] A solution to the above-mentioned challenges may involve receiving a handwritten signature or handwritten sample from a person, extracting at least one of an image of the handwritten signature or handwritten sample and a trajectory of the handwritten signature or handwritten sample from the received handwritten signature or handwritten sample, accessing a reference signature or handwritten sample, extracting at least one of a reference image and a reference trajectory from the accessible reference signature or handwritten sample, comparing the image of the handwritten signature or handwritten sample and / or the trajectory of the handwritten signature or handwritten sample to the reference image and / or the reference trajectory, and determining whether the handwritten signature or handwritten sample is authentic based on the comparison.

[0021] Examples in this disclosure include a method for encoding a genuine signature and / or handwritten sample directly onto a writable document, for example, directly onto a check or contract, with the ability to automatically extract and decode the genuine signature and / or handwritten sample from an image of the writable document, for example, a check or contract, during the verification process. Accordingly, the encoded genuine signature and / or handwritten sample can be used as a reliable reference point for a signature and handwriting verification algorithm. handwritten signatures improve the accuracy and reliability of the verification algorithm in detecting fraudulent behavior, such as check forgery. The examples in this disclosure significantly mitigate the risk of fraudulent activity by incorporating an authentic handwritten signature and / or sample into the verification process, thereby preserving the integrity of financial transactions and legal documents.

[0022] Another solution involves using neural networks, which are sets of weighted functions where the output of one function or weighted layer is the input of another function or weighted layer. In the case of a Siamese neural network, the neural subnetworks of the Siamese neural network have shared weights. Neural networks can be trained on known datasets of signatures or handwritten samples, and the trained neural networks can then be used on unknown datasets of signatures or handwritten samples.For example, neural networks can be trained on a number of known signatures or handwritten samples, as well as on a number of known forgeries, so that when presented with an unknown signature or handwritten sample, it is possible to determine with a certain degree of confidence whether the unknown signature or handwritten sample is in fact authentic or a forgery.

[0023] Encoding an image of a signature or biometric characteristics of a signature (trajectory) onto writable documents such as bank checks for additional signature verification

[0024] Examples of this disclosure include the use of authentic signature samples, for example, from an account holder, which are coded or embedded in a document or writable medium such as a bank check, as a source of reference data for further signature verification. In the case of financial institutions such as banks, bank checks are susceptible to signature fraud, which can be detected if a signature on a check can be compared in real time to an authentic signature belonging to the account holder. In most current banks, images of cards or reference signatures on historical archived checks whose authenticity has been proven are used as a reference.However, many transactions lack access to historical data, for example, ATM transactions or transactions at a bank branch that does not have access to an archived authentic reference signature, making signature verification difficult. Examples of implementations of this disclosure include a process of encoding an account holder's signatures directly onto the check, which allows immediate access to the reference signature that is encoded or embedded on the check. The check is examined to compare it to the signature written on the same check, at the point of sale, at an ATM, at a terminal, and at other locations or facilities that could be exploited by fraudulent check issuers. Encoding or embedding a signature on a check renders the signature inaccessible and virtually invisible to a potential criminal or forger, thus reducing or eliminating the possibility of falsifying a check and writing a signature different from that of the intended recipient, for example, to deposit the check into an account other than that of the intended recipient.

[0025] Determining the authenticity of a signature on a writable document, for example a check, involves the presence of a reference signature on the same writable document or check. The reference signature may be captured from an original reference signature medium, the medium being, for example, the writable document, a check, or a signature pad, or may be provided in the form of a trajectory. The trajectory characteristics may include at least one of a sequence of coordinates representing locations of a writing element, a timestamp corresponding to each location, the pressure applied by the writing element, and the tilt of the writing element.The authentic signature and / or trajectory reference can be embedded in the writable document in a coded form, making the authentic signature and / or trajectory reference significantly difficult to copy and significantly unusable for illicit purposes.

[0026] Various examples in this disclosure involve an initial operation of accessing a reference signature or a reference handwriting sample on a medium, and extracting either a reference image or a reference trajectory from the reference signature or the reference handwriting sample taken from the medium. The medium may be, for example, a piece of paper, a typing pad, or some other system or method for collecting, receiving, or gathering a signature or handwriting sample. In various cases, the reference signature may have been previously provided by the individual and may be retained as a reference for authenticating signatures provided subsequently.Access to the reference signature may involve, for example, retrieving a scan of a reference card completed by the account holder, archived historical checks bearing the account holder's authentic signature, or previously collected biometric signature characteristics belonging to the account holder. Extraction of the reference signature or reference handwriting sample from the medium can be performed using neural networks.

[0027] In various examples, the encoding of the authentic reference signature of the account holder on the writable document involves the transformation of the signature of A reference signature, for example, can be printed, scanned, embedded, affixed, or otherwise associated with the writable document. Consequently, the reference signature extracted from the medium can undergo transformation by a second trained neural network to generate integration vectors. The reference signature, in the form of one or more integration vectors, can be encoded using one of many methods, including, but not limited to, a quick response (QR) code, a barcode, a stacked linear barcode (PDF417), or a radio-frequency identification (RFID) tag, among others, and can be embedded on the front or back of the writable document, such as a check. In one example, when a check belonging to a person is issued to that person, the person's reference signature can be embedded on the check.

[0028] The written signature presented for verification is generally located on the writable document, for example, a check or contract, and can be extracted from the writable document for comparison with the reference signature, which is encoded or embedded on the same writable document. In some examples, the extraction of the written signature from the writable document can be performed using a neural network. Consequently, the written signature can be transformed by a neural network trained to extract and encode the written signature by generating integration vectors. The neural network used for the transformation of the written signature can be the same as the neural network used for the transformation of the reference signature.For example, the neural networks used for integrating the reference signature and for integrating the written signature can be Siamese neural networks.

[0029] During the verification process, the reference signature embedded on the front or back of the writable document is extracted and decoded to allow its comparison with the written signature. For example, the transformed reference signature and the transformed written signature can both then be provided to a third neural network, which can be trained to compare the integration vectors of the transformed reference signature and the transformed written signature. In one example, the third neural network can generate a score indicating the degree of confidence that the reference signature and the written signature belong to the same individual or to different individuals.

[0030] Encoding an image of a sample of handwriting on bank cheques in order to make the image available for the detection of fraudulent cheques

[0031] Similar to the above example of signature comparison, in this example, an authentic handwritten sample that is coded in a medium such as a check, contract or other document requiring handwriting verification, can also be a source of reference data for fraud detection. Currently, check fraud can be detected if handwriting on a check is compared to one or more samples of authentic handwriting belonging to the account holder. Generally, handwriting preserved on a check in fields other than those that have been whitened and altered (e.g., the payee's name, the amount) can be used as a reference handwriting sample. When archived checks containing the person's handwriting and whose authenticity is known are available, these archived checks can also be used as reference handwriting sources. However, if access to historical data is unavailable, verifying handwriting becomes difficult.The process of encoding a handwritten sample from an account holder onto a check, for example, can allow immediate access to reference data at a point of sale, ATM, terminal, or other location or situation that might be conducive to the use of fraudulent checks. Encoding or embedding a handwritten sample from the account holder onto a check renders the handwritten sample essentially inaccessible and virtually invisible to potential criminals, thereby reducing or eliminating the possibility of copying or imitating the account holder's authentic handwriting for criminal purposes, for example.

[0032] Determining the authenticity of handwriting on a written document requires the presence of a reference handwriting sample. In some disclosure examples, the reference handwriting is embedded in coded form or is coded on the same document, for example, a check. The reference handwriting sample may be captured from an original reference handwriting medium, such as an archived check image or an image of another document, or provided in the form of a reference path. Reference path features may include one or more coordinate sequences representing locations of a writing instrument, which may be a pen or other writing tool, a timestamp corresponding to each location of the writing instrument on the document, the pressure applied by the person via the writing instrument, and the tilt of the writing instrument.The reference trajectory characteristics can then be embedded in the writable document in coded form, making the reference trajectory characteristics substantially difficult to copy and substantially unusable for unlawful or criminal purposes.

[0033] According to various examples, an initial operation may involve accessing one or more handwriting reference samples and extracting either a reference image or a reference trajectory from the reference sample(s) on the reference handwriting medium. In various cases, The reference handwriting sample may have been previously provided by the individual and retained as a reference for authenticating subsequently submitted documents that are filled out in handwriting, or it may be captured from archived documents (checks) that have been proven authentic. Accessing the reference handwriting sample file may involve retrieving a scan of a reference document filled out by the account holder, archived historical checks bearing the account holder's authentic handwriting, or previously collected biometric handwriting characteristics belonging to the account holder. Extracting the reference handwriting from the original medium can be accomplished using, for example, neural networks. The original medium here refers to the medium on which the reference handwriting or check was captured.The original medium could be, for example, a sheet of paper, a notepad, a scan, a photograph, and the like.

[0034] In certain examples, when the reference handwriting is extracted from the original medium, the reference handwriting can undergo a transformation by a second neural network that is trained to generate integration vectors. As a result, the reference handwriting, in the form of integration vectors, can be encoded using one of many methods, including, but not limited to, a QR code, a barcode, a PDF417, or an RFID, and embedded on the front or back of the handwriting document, for example, on the front or back of a check or a contract.

[0035] In other examples, the handwriting or signature to be authenticated, which is generally located on the written document or check, for example, is extracted from the written document or check. For example, the handwriting or signature can be extracted from the written document using neural networks.

[0036] Next, the handwriting or signature to be authenticated is transformed to generate integration vectors. For example, the handwriting or signature to be authenticated is transformed by a trained neural network to generate the integration vectors. The neural network used for the transformation of the handwriting or signature to be authenticated may be the same as the neural network used for the transformation of the reference handwriting. For example, the neural networks used to integrate the reference handwriting and to integrate the handwriting or signature to be authenticated may be, or may include, Siamese neural networks, for example. During the verification process, the reference handwriting embedded on the front or back of the written document is extracted and decoded. The written document may be decoded into an integration format.

[0037] When the reference handwriting or signature and the handwriting or signature to be authenticated are both integrated as integration vectors, both integration vectors can be provided to a third neural network. In some examples, the third neural network can be trained to compare the integration vectors and to generate a score indicating the degree of confidence that the reference handwriting or signature and the handwriting or signature to be authenticated belong to the same individual or to different individuals.

[0038] Figure 1 is a diagram illustrating a method and system for detecting Forgery or counterfeiting of a writable document, according to various aspects. In [Fig. 1], in system 100, a reference handwriting sample 110 can be provided, for example, at a handwriting receiving device 120 by a person (not shown). In various aspects, the reference handwriting sample 110 may be or include a reference signature or a reference handwriting sample 110, and may be captured on the handwriting receiving device 120, for example, via a pen, stylus, other writing instrument, the person's finger, or another method of capturing a handwriting sample. In another example, the reference handwriting sample 110 may be captured as an image, such as an image of a pre-existing handwriting sample. The reference handwriting sample 110 may also be captured as a photograph or a scan.For example, an identity document may be provided as a scan, and from the scan of the identity document, an image of the handwritten sample or reference signature 110 may be extracted. In one example, the writing receiving device 120 may be, or may include, for example, a sheet or piece of paper, a signature pad, or a signature screen or display, such as a handheld device, a smartphone, a tablet, or another writing receiving device. The writing receiving device 120 may also be a terminal comprising a writing receiving surface, a video camera, a still camera, and / or a scanning device.

[0039] In various examples, the writing sample or reference signature 110 can be encrypted, for example, using encryption software or an encryption engine 130. For example, the writing sample or reference signature 110 can also be compressed using compression software or an engine 140, for example. In various aspects, and as discussed in more detail below, the encrypted and compressed writing sample or reference signature 110 can be saved in a data repository 150, such as memory, for example. The system 100 can also include an integration device 160. For example, the device The integration device 160 may be, or may include, a printing or encoding device 160 configured to print or encode the writing sample or reference signature 110 onto it, for example, a writable document. The writable document, such as a check, may correspond to the writable document 200 described in more detail below. Although the integration device 160 is shown as being directly connected to the other elements of the system 100, the integration device 160 may be separate from the other elements of the system 100 and may be used to integrate the writing sample or reference signature 110 by accessing the writing sample or reference signature 110 stored in a data repository such as the data repository 150 at the same time or at a different time than the writing sample or reference signature 110 is stored on the data repository 150.Furthermore, the integration device 160 can be remote from the data repository 150, and can access the data repository, for example via a cloud system 155, the Internet, an intranet or another remote access system or device.

[0040] Figure 2 illustrates a writable document into which a handwriting sample is embedded, in accordance with various aspects of this disclosure. In Figure 2, the writable document 200 includes one or more handwriting segments 210a-210c. In the example shown in Figure 2, the writable document 200 is a check. However, the writable document 200 could be another document or written instrument for which verification of a handwriting segment, such as a signature, is relevant to the validation of the document. For example, the writable document could be a contract, a will, a promissory note, an agreement, or other legal document requiring validation of a person's signature or handwriting. In various examples, the writable document 200 includes an embedded portion 220.

[0041] For example, the embedded portion 220 may be, or may include, an encrypted and compressed reference handwriting sample from the same person who is to write on the writable document 200. For example, the compressed and encrypted handwriting sample embedded in the embedded portion 220 is from the same person who is supposed to have signed or written any one or more of the handwriting segments 210a-210c. In the case of a check 200, the embedded portion 220 may be, or may include, an encrypted and compressed reference handwriting sample or signature from the same person who is to sign the check 200 in portion 210b. In the example shown in [Fig. 2], the embedded portion 220 may be, or may include, an encrypted and compressed reference handwriting sample or signature from John Brown. In one example, the integrated part 220 can be or may include at least one of the following: a printed label, a printed barcode, a code printed quick response, a magnetic stripe and a radio frequency identification tag.

[0042] Figure 3 is a diagram illustrating a system for receiving a handwriting sample or signature to be authenticated, in accordance with various aspects of this disclosure. In Figure 3, the system 300 includes a receiving device 310. For example, the receiving device 310 can be configured to receive the writable document 320, which may be similar to the writable document 200 shown in Figure 2 and which includes a handwritten segment or signature, such as signature 210b also shown in Figure 2. In some examples, the writable document 320 may be, or may include, a legal document or instrument, such as the check 200 shown in Figure 2, a will, a contract, a promissory note, an agreement, or some other document or instrument whose validity or use depends on the validity of a handwritten signature or signature affixed to it.The receiving device 310 can be coupled to an updatable data repository and a computer device (not shown) which may be similar to the data repository and computer device described below with reference to [Fig.6].

[0043] In various examples, the writable document 320 received on the receiving device 310 may include a handwritten portion or a signature 325 to be encrypted, for example, via encryption software or an encryption engine 330. For example, the encrypted handwritten portion or signature 325 received on the receiving device 310 as part of the writable document 320 may also be compressed, for example, via a compression engine or software 340. In various aspects, and as discussed in more detail below, the compressed and encrypted writable document 320 may be saved in a data repository, for example, such as memory (not shown). The receiving device 310 may also be functionally coupled to an updatable data repository and a computing device not shown in [Fig. 3] but described in more detail in [Fig. 6].

[0044] Figure 4 is a diagram illustrating a system for detecting counterfeiting or falsification of writable documents, in accordance with various aspects of this disclosure. In Figure 4, the system 400 includes an embedded handwriting sample or reference signature 410, for example, such as the handwriting sample or reference signature 110 illustrated in Figure 1. For example, the reference handwriting sample 410 may be encrypted and compressed and embedded on a writable document. With reference to Figure 2, the embedded encrypted and compressed reference handwriting sample 410 may correspond to the embedded portion 220 on the writable document 200. The system 400 also includes a handwritten segment 420 that is on or forms part of the writable document. For example, the The handwritten segment 420 can also be encrypted and compressed before being incorporated into the writable document. Referring to [Fig. 2], the handwritten segment 420 can correspond to the handwritten segment 210b on the writable document 200.

[0045] The system may further include a software engine or processor 430 configured to access and decrypt both the encrypted and compressed handwriting sample 410 and the encrypted and compressed handwritten segment 420. For example, the software engine may be, or may include, one or more neural networks. In another example, a neural network may be, or may include, a Siamese neural network. In various examples, the software engine 430 can access and decrypt both the handwriting sample 410 and the handwritten segment 420. The software engine 430 can also compare the decrypted handwriting sample 410 and the handwritten segment 420, and, based on the comparison, determine whether the handwriting sample 410 and the handwritten segment 420 were written by the same person.

[0046] Figure 5 is a flowchart describing a method for detecting counterfeiting of a writable document, in accordance with various aspects. For convenience only, the method 500 is described using at least the example of the system 600 described below. However, it is understood that the method 500 can be implemented by any suitable system.

[0047] Operation 510 involves encrypting a reference handwriting sample, for example, from a person. For instance, the person may enter their reference handwriting sample, such as a signature, onto a receiving device. The receiving device may be a sheet of paper, an electronic pad, or another device configured to receive a person's handwriting or signature. When the reference handwriting sample is entered, it may be scanned and / or stored in memory in digital form. In various aspects, the encryption of the reference handwriting sample involves encrypting one of the following: an image of the reference handwriting sample and a biometric characteristic of the reference handwriting sample.In the case of a signature, the encryption of the reference handwriting sample involves encrypting one of the signature's images and a biometric characteristic. For example, the encryption of the reference handwriting sample can be performed using encryption software. Alternatively, the handwriting sample can be encrypted using a neural network, such as a Siamese neural network.

[0048] Operation 520 involves compressing the encrypted reference writing sample. For example, the compression of the encrypted reference writing sample can be performed by compression software. For example, the compression of The encrypted reference writing sample can be produced via a neural network such as a Siamese neural network, for example.

[0049] Operation 530 involves embedding the encrypted and compressed reference handwriting sample into the writable document. In the case of a bank check, operation 530 involves embedding or printing the encrypted and compressed reference handwriting sample onto a portion of the check. With reference to [Fig. 2], operation 520 involves embedding the embedded portion 220 onto the check 200. In various aspects, embedding the encrypted and compressed reference handwriting sample into the writable document can be accomplished by generating on the writable document at least one of the following: a readable tag, such as a machine-readable or human-readable tag, a printed label, a printed barcode, a printed quick-response code, a magnetic stripe, a radio-frequency identification tag, and the like.

[0050] Operation 540 involves receiving the writable document, the writable document having a handwritten segment. With reference to [Fig. 2], operation 540 involves receiving the writable document or check 200 which has the handwritten segment of any of the handwritten segments 210a-210c bearing the signature 210b. For example, when the writable document is a check, operation 540 involves receiving the check, and the check has a signature of the person as well as the reference handwriting sample embedded therein. In various aspects, the receipt of the writable document can be accomplished by scanning at least the handwritten segment of the writable document, for example, the signature. In another example, the receipt of the writable document involves receiving the writable document at one of two points, such as an automated teller machine and a receiving institution.For example, the receiving institution could be a bank, a car dealership, a law firm, a notary, or any other institution likely to receive a check or legal instrument whose validity depends on verifying a signature on it.

[0051] Operation 550 involves encrypting the handwritten segment of the received writable document. With reference to [Fig. 2], operation 550 involves encrypting any one of the handwritten segments 210a-210c containing the signature 210b. In one example of operation 550, the encryption of the handwritten segment included in the writable document can be achieved by encrypting an image of the handwritten segment, for example, an image of a signature, or a biometric characteristic of the handwritten segment, for example, a speed, an angle, or another characteristic of the signature.

[0052] Operation 560 involves comparing the encrypted reference handwriting sample to the encrypted handwritten segment. With reference to [Fig. 2], operation 560 involves comparing the encrypted integrated part 220, which includes the sample of handwriting or the signature referencing a ciphertext handwritten segment such as any one or more of the 210a-210c handwritten segments. For example, the comparison of the ciphertext handwriting sample to the ciphertext handwritten segment can be performed using a software engine such as a neural network. An example of a neural network is a Siamese neural network.

[0053] Operation 570 involves assessing the probability that the ciphered reference handwriting sample and the ciphered handwritten segment were written by the same person based on the comparison. For example, assessing the probability that the ciphered handwriting sample and the ciphered handwritten segment were written by the same person may involve generating a similarity factor between the ciphered handwriting sample and the ciphered handwritten segment. For example, operation 570 may involve establishing a similarity factor threshold, and when the generated similarity factor is equal to or greater than the established similarity factor threshold, the ciphered handwriting sample and the ciphered handwritten segment are considered to have been written by the same person.

[0054] Figure 6 represents a schematic diagram of a computing device, according to various aspects. In the illustrated example, the computing device 600 may include a bus 602 or some other communication mechanism of similar function for communicating information, and one or more processing elements 604 (collectively referred to as a processing element 604) coupled to the bus 602 for processing the information. As those versed in the art will understand, the processing element 604 may include a plurality of elements or processing cores, which may be grouped into a single processor or in a distributed arrangement. Furthermore, a plurality of virtual processing elements 604 may be included in the computing device 600 to provide, for example, the compression, encryption, and comparison operations or the process 500 illustrated and disclosed above.

[0055] The computing device 600 may also include one or more volatile memories 606, which may, for example, include one or more random access memories (RAM) or other dynamic memory components, coupled to one or more buses 602 for use by at least one processing element 604. The computing device 600 may further include one or more non-volatile static memories 608, such as read-only memories (ROM) or other static memory components, coupled to the buses 602 to store information and instructions for use by at least one processing element 604. A storage component 610, such as a storage disk or a storage memory, may be provided to store information and instructions for use by at least one processing element 604. As will be understood, the computing device 600 may include a distributed storage component 612, such as a network disk or other storage resource available to the computing device 600.

[0056] The computer device 600 can be coupled to one or more displays 614 to display information to a user. One or more optional user input devices 616, such as a keyboard and / or a touchscreen, can be coupled to the bus 602 to communicate information and control selections to at least one processing element 604. An optional cursor control or graphics input device 618, such as a mouse, trackball, or cursor direction keys, can be used to communicate graphical user interface information and control selections to at least one processing element.The computer device 600 may further include an input / output (I / O) component, such as a serial connection, a digital connection, a network connection or other input / output component enabling intercommunication with other computer components and the various components of systems 100, 300 and 400 or of process 500 illustrated and disclosed above.

[0057] In various embodiments, the computer device 600 can be connected to one or more other computer systems via a network to form a networked system. These networks may, for example, include one or more private or public networks, such as the Internet. In the networked system, one or more computer systems can store and serve data to other computer systems. The one or more computer systems that store and serve data may be called servers or the cloud in a cloud computing scenario. The one or more computer systems may include one or more web servers, for example. The other computer systems that send and receive data to and from the servers or the cloud may be called client devices or the cloud, for example.Various compression, encryption and comparison operations, or the 500 process illustrated and disclosed above, can be supported by the operation of distributed computer systems.

[0058] The computer device 600 can be operational for controlling various operations of the process 500 illustrated above via a communication device, such as the communication device 620, for example, and for manipulating data, such as encrypted and compressed data. In some examples, analysis results are provided by the computer device 600 in response to the execution of instructions contained in a memory 606 or 608 by at least one processing element 604 and the performance by the latter of operations on the received data. The execution of instructions contained in a memory 606 and / or 608 by at least one processing element 604 can make the process 500 operational for carrying out the processes described herein.

[0059] The term "computer-readable medium" as used herein refers to any medium that participates in providing instructions to the processing element 604 for execution. Such a medium can take many forms, including, but not limited to, non-volatile media, volatile media, and transmission media. Non-volatile media include, for example, optical or magnetic disks, such as disk storage 610. Volatile media include dynamic memory, such as memory 606. Transmission media include coaxial cables, copper wires, and optical fibers, including wires that carry the 602 bus.

[0060] Common forms of computer-readable media or computer program products include, for example, a floppy disk, a flexible disk, a hard disk, a magnetic tape, or any other magnetic medium, a CD-ROM, a digital video disc (DVD), a Blu-ray disc, any other optical medium, a USB flash drive, a memory card, RAM, PROM and EPROM, FLASH-EPROM, any other memory chip or cartridge, or any other tangible medium from which a computer can read.

[0061] Various forms of computer-readable media can be used to transport one or more sequences of one or more instructions to the processing element 604 for execution. For example, the instructions can initially be transported by the magnetic disk of a remote computer. The remote computer can load the instructions into its dynamic memory and send them over a telephone line using a modem. A local modem at the computing device 600 can receive the data over the telephone line and use an infrared transmitter to convert the data into an infrared signal. An infrared detector coupled to the bus 602 can receive the data carried in the infrared signal and place the data on the bus 602. The bus 602 carries the data to memory 606, from which the processing element 604 retrieves and executes the instructions.Instructions received by memory 606 and / or memory 608 may optionally be stored on storage device 610, either before or after their execution by processing element 604.

[0062] According to various embodiments, the instructions to be executed by a processing element to implement a process are stored on a computer-readable medium. The computer-readable medium may be a device that stores digital information. For example, a computer-readable medium may include a compact disc read-only memory (CD-ROM) such as is known in the art for storing software. The computer-readable medium is accessed by a processor adapted to execute instructions configured to be executed.

[0063] This disclosure has described certain examples of current technology with reference to the accompanying drawings, in which only some of the possible examples have been shown. Other aspects, however, can be embodied in many different forms and should not be interpreted as being limited to the examples shown herein. Rather, these examples have been provided to make this disclosure exhaustive and complete and to fully convey the scope of possible examples to those skilled in the art.

[0064] Although specific examples have been described herein, the scope of the technology is not limited to these specific examples. A person skilled in the art will recognize other examples or improvements that fall within the scope of this technology. Accordingly, the specific structure, actions, or supports are disclosed only as illustrative examples. Examples according to the technology may also combine elements or components of those disclosed generally but not expressly illustrated in combination, unless otherwise stated herein. The scope of the technology is defined by the following claims and all their equivalents.

Claims

Demands

1. A method for detecting counterfeiting of a writable document, the method comprising the following steps: encrypting a reference handwriting sample; compressing the reference handwriting sample; embedding the reference handwriting sample in the writable document; receiving the writable document, the writable document having a handwritten segment; encrypting the handwritten segment of the received writable document; comparing the encrypted reference handwriting sample with the encrypted handwritten segment; and evaluating a probability that the encrypted reference handwriting sample and the encrypted handwritten segment were written by the same person based on the comparison.

2. A method according to claim 1, wherein the encryption of the reference handwriting sample comprises the encryption of one of an image of the reference handwriting sample and a biometric feature of the reference handwriting sample.

3. A method according to claim 1, wherein the encryption of the reference handwriting sample comprises the encryption of one of an image of a signature and a biometric feature of the signature.

4. A method according to claim 1, wherein at least one of the encryption, compression and comparison includes the use of a neural network.

5. A method according to claim 1, wherein the integration of the encrypted and compressed reference writing sample into the writable document comprises the generation on the writable document of at least one of: a readable tag; a printed label; a printed barcode; a printed quick response code; a magnetic stripe; and a radio frequency identification tag.

6. Method according to claim 1, wherein the receipt of the writable document includes the receipt of the writable document at one of a bank automated teller machine and a receiving institution.

7. A method according to claim 1, wherein the receipt of the writable document includes the scanning of at least the handwritten segment of the received writable document.

8. A method according to claim 1, wherein the writable document comprises one of a check and a legal instrument.

9. Method according to claim 1, wherein the encryption of the handwritten segment included in the writable document comprises the encryption of one of an image of the handwritten segment and a biometric feature of the handwritten segment.

10. A method according to claim 1, wherein the evaluation of the probability that the ciphered reference handwriting sample and the ciphered handwritten segment were written by the same person includes the generation of a similarity factor between the ciphered reference handwriting sample and the ciphered handwritten segment.

11. A method according to claim 10, further comprising the following steps: establishing a similarity factor threshold; wherein when the generated similarity factor is equal to or greater than the established similarity factor threshold, the ciphered reference handwriting sample and the ciphered handwritten segment are considered to have been written by the same person.

12. A system for detecting counterfeiting of writable documents comprising: an updatable data repository; an integration device; and a computing device operationally coupled to the updatable data repository and the integration device, the computing device comprising a processor and memory, the memory storing instructions which, when executed by the processor, perform a set of operations comprising the following: via the processor, encrypting a reference writing sample; via the processor, compressing the reference writing sample; and via the integration device, integrate the reference writing sample into the writable document.

13. System according to claim 12, wherein the processor comprises a neural network.

14. System according to claim 12, wherein the instruction set includes encrypting the reference handwriting sample by encrypting one from an image of the reference handwriting sample and a biometric feature of the reference handwriting sample.

15. System according to claim 12, wherein the instruction set includes encrypting the reference handwriting sample by encrypting one from an image of a signature and a biometric feature of the signature.

16. System according to claim 12, wherein the set of operations comprises integrating the reference writing sample into the writable document by generating at least one of a readable tag, a printed label, a printed barcode, a printed quick response code and a radio frequency identification tag on the writable document.

17. A system for detecting counterfeit writable documents comprising: a receiving device; an updatable data repository functionally coupled to the receiving device; and a computing device operationally coupled to the receiving device and the updatable data repository, the computing device comprising memory, the memory storing instructions which, when executed by the processor, perform a set of operations comprising the following: receiving the writable document on the receiving device, the writable document having a handwritten segment and a reference writing sample embedded therein; via the processor, encrypting the handwritten segment of the writable document; via the processor, comparing the encrypted reference writing sample to the encrypted handwritten segment; and via the processor, evaluate a probability that the encrypted reference handwriting sample and the encrypted handwritten segment were written by the same person based on the comparison.

18. System according to claim 17, wherein the processor comprises a neural network.

19. System according to claim 17, wherein the writable document comprises one of a check and a legal instrument.

20. System according to claim 17, wherein the receiving device comprises one of: a bank automated teller machine; and a reading machine at a receiving institution.

21. System according to claim 17, wherein the set of operations includes receiving the writable document on the receiving device by scanning at least the handwritten segment of the writable document received on the receiving device.

22. System according to claim 17, wherein the set of operations includes the encryption of the handwritten segment included in the writable document by encrypting one of an image of the handwritten segment and a biometric feature of the handwritten segment.

23. System according to claim 17, wherein the set of operations includes evaluating the probability by generating a similarity factor between the ciphered reference handwriting sample and the ciphered handwritten segment.

24. System according to claim 23, wherein the set of operations further comprises the following: establishing a similarity factor threshold; wherein when the generated similarity factor is equal to or greater than the established similarity factor threshold, the ciphered reference handwriting sample and the ciphered handwritten segment are deemed to have been written by the same person.

25. A system for detecting counterfeit writable documents comprising: a receiving device; an integration device; at least one updatable data repository functionally coupled to the receiving device and the integration device; at least one computer device operationally coupled to the receiving device and at least one updatable data repository, the at least one computer device comprising at least one processor and at least one memory, the memory storing instructions which, when executed by the at least one processor, perform a set of operations comprising the following: using at least one processor, encrypt a reference writing sample; via at least one processor, compress the reference write sample; via the integration device, integrate the reference writing sample into the writable document; receive the writable document on the receiving device, the writable document containing a handwritten segment; using at least one processor, encrypt the handwritten segment of the writable document; using at least one processor, compare the encrypted reference handwriting sample to the encrypted handwritten segment; and using at least one processor, evaluate a probability that the encrypted reference handwriting sample and the encrypted handwritten segment were written by the same person based on the comparison.

26. System according to claim 25, wherein at least one of the reference handwriting sample and the handwritten segment includes a signature.

27. ​​System according to claim 25, wherein the processor comprises a neural network.

28. System according to claim 25, wherein the instruction set includes encrypting the reference handwriting sample by encrypting one from an image of a signature and a biometric feature of the signature.

29. System according to claim 25, wherein the instruction set includes encrypting the reference handwriting sample by encrypting one of an image of the reference handwriting sample and a biometric feature of the reference handwriting sample.

30. System according to claim 26, wherein the set of operations includes encrypting the handwritten segment of the writable document by encrypting one of an image of the handwritten segment and a biometric feature of the handwritten segment.

Images