EP2909817A1

EP2909817A1 - Methods, devices and identification document for a person or an animal

Info

Publication number: EP2909817A1
Application number: EP13789856.5A
Authority: EP
Inventors: Zbigniew Sagan; Jean-Pierre Massicot; Alain Foucou
Original assignee: Advanced Track and Trace SA
Current assignee: Advanced Track and Trace SA
Priority date: 2012-10-16
Filing date: 2013-10-16
Publication date: 2015-08-26
Also published as: US20150266331A1; WO2014060700A1; FR2996943A1; FR2996943B1; US10112431B2

Abstract

The invention relates to a device (10) for producing a document (40) carrying the identification of a person or an animal, said device comprising: a means (190) for determining printing conditions of said document; a means (105) for obtaining the biometric data of said person or said animal; a means (110) for encoding the biometric data in order to form a digital identification image comprising redundancies; a means (195) for determining physical characteristics of cells of at least one form, according to the printing conditions, in such a way that the proportion of cells printed with a printing error resulting exclusively from printing problems is higher than a first predetermined value and lower than a second predetermined value; an encoding means (115) for forming, within said form, an image for the authentication of said document, said authentication image comprising elementary cells, the content of which represents the encoded authentication image; and a means (120) for printing the authentication image and the identification image on the document, implementing said printing conditions, in order to generate random errors in the printed elementary cells of the authentication image, said form being suitable for allowing the detection of a copy modifying the appearance of a plurality of said elementary cells.

Description

METHODS, DEVICES AND DOCUMENT FOR IDENTIFYING A PERSON OR A PERSON

AN IMAL

The present invention relates to a method, a device and a document identifying a person or an animal. It applies, in particular, to the production of identity documents and the verification of authenticity and integrity of identity documents.

Many types of identity documents are known. However, although they are difficult to falsify, techniques exist to modify the biometric information they carry, be it a photograph, a fingerprint, the size, the hair color or the eyes of their wearer.

In addition, the verification of these elements is difficult because, once modified, they can correspond to a new holder of the identity document.

The present invention aims to remedy these disadvantages.

For this purpose, according to a first aspect, the present invention relates to a device according to claim 1.

Thanks to these provisions, the biometric data of the document holder are coded and particularly difficult to falsify. In addition, the printed authentication image has, from its first printing on the document, a significant amount of random errors which require that a copy, made with the same printing means as the original document, of the Printed identification image would have even more errors, the total number of errors of the copy to detect that it is a copy. The redundancies present in the identification image make it possible to correct the printing and / or reading defects and the marks of wear or erasure of the identification image.

In embodiments, the biometric data obtaining means is configured to obtain a photograph of the head of the person or animal and the biometric data encoding means is configured to encode the photograph to form the image. Identification.

Thanks to these provisions, a decoder can display the photograph of the person or animal and verify his identity.

In embodiments, the biometric data encoding means is configured to encode the biometric data into at least one hundred thousand elementary cells.

Thanks to these arrangements, a compressed photograph in four kilobytes can be represented with several redundancies. This photograph can be very precise.

In embodiments, the biometric data encoding means is configured to apply a redundancy factor of at least three.

Thanks to these provisions, detection and correction of printing errors or image capture and image wear or erasure can be achieved even with a lot of lost data.

In embodiments, each encoding means is configured to form an image capable of being printed at a resolution of 600 dots per inch while retaining its reading capabilities. In embodiments, the biometric data obtaining means is configured to obtain data representative of at least one fingerprint of the person and the biometric data encoding means is configured to encode the data representative of the fingerprint. digitalis.

Thanks to these provisions, the identity of the identity document holder can be verified by comparing each fingerprint data represented by the identification image with the corresponding data of a fingerprint provided during the identification.

In embodiments, the device that is the subject of the invention comprises means for forming, on the document, at least one filing surface of a fingerprint.

Thanks to these provisions, the capture of at least one fingerprint of the holder of the identity document is made directly on the identity document after pressing the finger concerned on the corresponding deposition surface.

In embodiments, the encoding means of an authentication image is configured to form a suitable image, when printed by the printing means, to comprise at least one tenth of elementary cells whose content is wrong.

With these provisions, the detection of the copy of the document is facilitated because it imposes an error rate of at least fifteen percent.

In embodiments, the encoding means of an authentication image is configured to encode at least one document holder identification information in the authentication image.

Thanks to these provisions, the authentication code, which can not be falsified, identifies the holder of the document, even if the identification image has been falsified.

In embodiments, the encoding means of an authentication image is configured to perform a permutation of elementary cells by implementing a hashing function, or condensate, implementing a hashing key.

Thanks to these provisions, the recovery of identification data, by a malicious person, is complicated.

In embodiments, at least one of the encoding means is configured to perform a permutation of elementary cells representative of a fingerprint by implementing a hashing function, or condensate, implementing a hashing key.

Thanks to these provisions, the recovery of fingerprint data, by a malicious person, is complicated.

Moreover, by implementing an inverse function of the hashing function, when reading the identification image thus produced, the error rate is reduced and thus the fingerprint is authenticated.

According to a second aspect, the present invention aims a device object of the claim

12.

According to a third aspect, the present invention provides a method according to claim 13. According to a fourth aspect, the present invention is directed to a method that is the subject of the claim In embodiments, the verification method includes the step of displaying the biometric data after a step of capturing biometric data of the document holder and a step of matching verification between the captured biometric data and the decoded biometric data.

Thanks to these provisions, the malicious people can not have access to the data represented by the identification image. The falsification of the document is thus made more difficult.

In addition, only the holder of the document can thus have access to data stored in the document.

According to a fifth aspect, the present invention relates to a document which is the object of the claim

16.

In embodiments, the object document of the invention comprises a deposition surface of at least one fingerprint.

Since the advantages, aims and particular characteristics of this verification device, these methods and this document are similar to those of the device for producing a document which is the subject of the present invention, they are not recalled here.

Other advantages, aims and particular features of the present invention will emerge from the description which follows, for an explanatory and non-limiting purpose, with reference to the appended drawings, in which:

FIG. 1 represents, schematically, particular embodiments of the device for producing an identity document and the verification device that is the subject of the present invention,

FIG. 2 represents, in the form of a logic diagram, steps of a particular embodiment of the method for producing an identity document that is the subject of the present invention; FIG. 3 represents, in the form of a logic diagram; , steps of a particular embodiment of the verification method that is the subject of the present invention,

FIG. 4 represents, schematically, an identity document that is the subject of the present invention and

FIG. 5 represents, in the form of a logic diagram, steps implemented to produce an authentication image.

As of now, we note that the figures are not to scale.

FIG. 1 shows a device 10 for producing a document 40 (an example of which is shown in FIG. 4) supporting identification data of a person or an animal. In the remainder of the description, such a document is called an "identity document". This document 40 takes, for example, the form of an identity card, a driver's license, a passport, a membership card, a medical card or a payment card.

The device 10 comprises:

means 1 90 for determining conditions for printing said document;

means 105 for obtaining biometric data representative of this person or this animal, means 1 for coding the biometric data to form a digital identification image comprising redundancies,

a means 195 for determining the physical characteristics of cells of at least one shape, as a function of the printing conditions, in such a way that the proportion of cells printed with a printing error originating exclusively from the print randomness is greater than at a first predetermined value and below a second predetermined value;

encoding means for forming an authentication image of the document, said authentication image having elementary cells and

means 120 for printing the authentication image and the identification image on the document 40 for generating random errors in the printed elementary cells of the authentication image.

The means 10 for obtaining biometric data representative of this person or this animal may comprise at least one sensor, for example a sensor of an image of the head of the future holder of the identity document or an impression sensor. digitalis. The means 105 may also include means for receiving data stored in a computer memory, for example a memory of a bank, a tax service or a social security organization, memory retaining biometric data, such as the size, color of hair or eyes of the future bearer of the identity document.

The means 1 for coding the biometric data to form a digital identification image implements a central unit 125 operating according to a program 130 whose instructions are stored in a memory 135. This program implements, for example, the steps illustrated in FIG. figure 2.

The coding means 1 for forming an authentication image of the document also implements the CPU 125 and the program 130.

The means 120 for printing the authentication image and the identification image on the document 40 is, for example consisting of a laser printer, inkjet or piezoelectric crystal.

The printing means 120 is configured to, at the resolution of the authentication image to be printed, generate random errors in the printed elementary cells of the authentication image, as explained with the details of the means 190 and 195 .

Thus, the biometric data of the document holder are encoded and particularly difficult to falsify. In addition, the printed authentication image has, from its first printing on the document, a significant amount of random errors which require that a copy, made with the same printing means as the original document, of the Printed identification image would have even more errors, the total number of errors of the copy to detect that it is a copy. The redundancies present in the identification image make it possible to correct the printing and / or reading defects and the marks of wear or erasure of the identification image.

As regards the means 190 and 195, in the rest of the description "printing error", a change in the appearance of a cell that modifies the interpretation of the information carried by this cell, during an analysis free from errors of reading or capture, for example microscopic. Note that if the cells often have binary values, the captured images are frequently in gray level, so we have a non-binary value associated with a cell; the latter can for example be interpreted as a probability on the original binary value of the cell.

If a gray level limit value is used to read to determine whether a cell is one color (eg white) or another (eg black), a printing error is an interpretation erroneous color of the cell of the original digital document.

The inventors have discovered that, when the proportion of printing error is greater than a predetermined value, the copy of the form made by implementing the same printing means as the original printing, or similar means, causes necessarily an additional proportion of errors making this copy detectable.

The inventors have also discovered that, depending on given constraints (such as a physical size or cell number constraint of the secure information matrix "MIS"), there is a proportion of optimal printing errors in term copy detection capability. This proportion of optimal printing error corresponds to a given cell size or print resolution depending on the printing medium.

Thus, contrary to prejudice, the highest print resolution is not necessarily, and rarely even, the best resolution in terms of copy detection capability.

Here, it is necessary to differentiate between the native print resolution of the printing means and the print resolution of the cells, which are each, in general, of a plurality of ink dots, each corresponding ink dot. to the native print resolution. Formally, you can not vary the print resolution of an MIS. Indeed, the majority of printing means print in binary (presence or absence of an ink dot) at a fixed resolution, and the gray or color levels are simulated by the different techniques of dithering. In the case of offset printing, this "native" resolution is determined by the resolution of the plate, which is, for example, 2,400 dots / inch (2,400 dpi). Thus, a grayscale image to be printed at 300 pixels / inch (300 dpi) would actually be printed in binary at 2,400 dpi, each pixel corresponding to approximately 8 x 8 points of the frame.

If one can not, generally, not vary the printing resolution, one can on the other hand vary the size in pixels of the cells of the M IS, so that a cell is represented by several printing points. Thus, one can for example represent a cell by a square block of 1 x 1, 2 x 2, 3 x 3, 4 x 4 or 5 x 5 pixels (non-square blocks are also possible), corresponding to resolutions of respectively 2,400, 1,200, 800, 600 and 480 cells / inch.

In embodiments, during the step of determining the physical characteristics of cells, the size of the cells to be printed is determined.

In embodiments, during the step of determining the physical characteristics of cells, a sub-part of the cells is determined, which sub-part is of a color uniform and variable to represent different values of an information, said sub-part being strictly less than said cell

In embodiments, the first predetermined value is greater than 5%, preferably 10% and even more preferably 15% and then 20%.

In embodiments, the second predetermined value is less than 30% and, preferably, 25%.

Before giving details of various particular embodiments of certain aspects of the present invention, definitions are given below which will be used in the description.

"Information matrix": this is a physical representation of a message, usually affixed to a solid surface (unlike watermarks or steganographies that modify the pixel values of a decor to be printed), readable by a machine (in English "machine-readable representation of information"). The definition of the information matrix includes, for example, 2D barcodes, one-dimensional barcodes and other information representation means that are less intrusive, such as "dataglyphs" of data) ;

"Cell": this is an element of the information matrix that represents an information unit;

"Document": this is any (physical) object carrying an information matrix; "Marking" or "printing": any process by which we move from a digital image (including an information matrix, a document ..) to its representation in the real world, this representation being generally made on a surface: this includes, non-exclusively, inkjet, laser, offset, thermal printing, as well as embossing, laser engraving, hologram generation. More complex processes, such as molding, in which the information matrix is first etched into the mold, then molded onto each object, are also included (note that a "molded" information matrix can be seen as having three dimensions in the physical world even if its digital representation has two of them.It should also be noted that many of the processes mentioned include several transformations, for example conventional offset printing (unlike the "computer-to-plate" offset), includes creating a film, said film providing for creating a plate, said plate being used in the printing, other methods also for printing information in the non-visible domain, or using frequencies at the same time. outside the visible spectrum, or to write the information inside the surface, etc., and

"Capture": any process by which we obtain a digital representation of the real world, including the digital representation of a physical document containing an information matrix.

Throughout the description that follows, we implement generally square shapes. However, the present invention is not limited to this type of form but extends, on the contrary, to all forms that can be printed. For example, forms made of M IS with different resolutions and ink charges, as set forth below, can be implemented, which would have the advantage, in particular, that at least one M IS corresponds to an optimum of resolution and an optimum of inking density.

Throughout the description, it implements a filling of the printed form that can be represented by a matrix of cells. However, the present invention is not limited to this type of form, but extends, on the contrary, to any filling by cells of identical or different shapes and sizes.

By way of introduction to the description of particular embodiments of the method and device of the present invention, it is recalled that the degradation of an information matrix has the consequence that the contents of certain cells may not be correctly decoded.

Each step of the creation of the information matrix is carried out for the purpose that the original message is readable without error, even if, and this is a desired effect, the initial reading of the information matrix is tainted with errors. In particular, one of the purposes of this information matrix creation is to use the number or the error rate of the encoded, replicated, permuted or scrambled messages to determine the authenticity of a brand of the information matrix. and therefore the document that carries it.

Indeed, the rate of this degradation can be adjusted according to the characteristics of the printing, so that the production of a copy leads to additional errors, resulting in a higher average error rate when reading. of a copy, only when reading an original.

To understand why a measurement of the error rate of the message can be sufficient to determine whether a document is an original or a copy, an analogy with the communication systems is useful. Indeed, the passage of the scrambled coded message to the information matrix that represents it is nothing but a modulation of the message, this modulation being defined as the process by which the message is transformed from its original form into a form. suitable for transmission on a channel. This communication channel, the information transmission medium that connects the source to the recipient and allows the routing of the message, differs depending on whether the captured information matrix is a captured original information matrix or an information matrix. copied captured. The communication channel can vary: one distinguishes the "communication channel of an original" and the "communication channel of a copy". This difference can be measured in terms of signal / noise ratio, this ratio being lower for a captured copied information matrix.

The coded message extracted from a captured copied information matrix will have more errors than the encoded message extracted from an original captured information matrix. The number or rate of errors detected are, in accordance with some aspects of the present invention, used to differentiate a copy of an original.

The communication channel of an original and the communication channel of a copy are advantageously described in terms of the subchannels that compose them, these being partly different in both cases. In the discussion below, each subchannel of the signal transmission channel, i.e. the information matrix, is an analog-to-digital or digital-to-analog transformation. The logic diagram illustrated in FIG. 5 exposes various steps for generating an information and marking matrix of a document, according to a particular embodiment of certain aspects of the method that is the subject of the present invention.

After starting, during a step 505, at least one marking or printing characteristic is received or measured during a step 510, for example the type of printing, the type of support, the type of ink used. Then, during a step 515, it is determined whether the area of the ISM or its cell number is set for the considered application or the client considered. During a step 520, the inking density corresponding to the marking / printing characteristics is determined, for example, by reading, in a database or a look-up table, the density corresponding to the printing characteristics. During a step 525, the size of the cells of the ISM is determined, for example, by reading, in a database or correspondence table, the size of cells corresponding to the printing characteristics. These correspondences are aimed at obtaining a good print quality and a proportion of printing errors between a first predetermined value and a second predetermined value, for example 5%, 10%, 15% or 20% for the first predetermined value and 25% or 30% for the second predetermined value.

Then, during a step 530, a message to be carried by a document is received, this message generally being a function of an identifier of the document, and, during a step 535, at least one secret key of encryption and / or scrambling.

The original message represents, for example, a designation of the document, the owner or owners of the intellectual property rights attached, a production order, a date of manufacture, a destination of the document, a manufacturing provider. It is constituted according to techniques known per se. The original message is represented in a predefined alphabet, for example in alphanumeric characters.

During a step 540, the message is encrypted, with a symmetric key or, preferably, with an asymmetric key, for example of the two-key type of public key infrastructure PKI (acronym for "public key infrastructure") to provide an encrypted message. Thus, to increase the level of security of the message, the message is encrypted or encrypted in such a way that a variation of a single binary message information, in encryption input, varies a large amount of binary information by encryption output.

The encryption generally operates on blocks of bits, of fixed size, for example 64 bits or 128 bits. The encryption algorithm DES (acronym for "standard data encryption" for standard data encryption) with a 56-bit key and a 64-bit message block size, the triple-DES, with a 168-bit key and a 64-bit message block size, and AES (advanced encryption standard), with a 128, 128 or 256-bit key and a 128-bit message block size , can be used because they are widely used, and recognized as being resistant to attack. However, many other block or sequential encryption algorithms can also be used. Note that the block cipher algorithms provide, in principle, encrypted messages of equal size to the initial message, provided that it is a multiple of the block size. During a step 545, the encrypted message is encoded to generate an encoded encrypted message. Preferentially, the encoding implements a convolutional coding, which is very fast to generate, the decoding being itself fast using, for example, the very well known method developed by Viterbi. If the convolutional coding used implements a generator polynomial of degree nine, and the code rate is two bits output for an input bit, a coding gain of seven dB will be obtained with respect to the same message that would simply be replicated. . This results in a much lower risk of decoding error. For a message to be encoded containing 128 bits, with the convolutional code previously described, there will be a 272-bit encoded message (there are two bits output for each of the 128 bits of the code and the eight bits belonging to the memory of the encoder for a generator polynomial of degree nine). Note however that many other types of encoding can be realized (arithmetic coding, turbo-code, ..) according to the same principle.

Preferably, this encoded encrypted message is then written in binary alphabet, that is to say that it is composed of "0" and "1".

During a step 550, the encrypted message encoded in a list of available cells is inserted into and replicated with an information matrix whose unavailable zones support synchronization, alignment or positioning symbols, or patterns of data. a location aid which, in embodiments, is determined from a secret key. The alignment patterns are, for example, 9 x 9 pixel matrices distributed periodically in the information matrix. The encoded encrypted message is thus replicated, or repeated, so that each binary information is represented several times, to correspond to the number of cells available in the information matrix. This replication, which is similar to a repetition or redundancy coding, makes it possible to significantly reduce the error rate of the encoded message that will be inputted to the convolutional code decoding algorithm. Errors not corrected by repetitions will be corrected by the convolutional code in most cases.

During steps 555 and 560, the replicated encoded encrypted message is scrambled, using so-called scrambling techniques, or scrambling, to provide a scrambled encoded encrypted message.

The jamming function, or scrambling, of the encoded replicated encrypted message preferably consists in successively applying a permutation, step 235, and a substitution, step 560, each depending on a second secret key, possibly identical to the first secret key, of binary values of the message. The substitution is preferably made from an "exclusive" function and a pseudo-random sequence.

Thus, the scrambling of the coded encrypted message is performed in a non-trivial way, by implementing a secret key, which can be a key identical to that used for the encryption of the message or a different key. Note that if the key is different, in particular embodiments, it can be calculated from a function of the key used for encryption.

The use of a secret key, both for the encryption of the message and for the scrambling of the encoded message, makes it possible to achieve a high level of security against counterfeits. By comparison, the existing methods of 2D bar code creation do not interfere with the encoded message, the counterfeiter can easily recreate an original information matrix after having decoding the message of the captured information matrix; even if the decoded message is encrypted, it does not need to decrypt the said message to recreate the information matrix identically.

The interference here preferably consists of a combination of permutation, step 555, and step 560, of the use of an "XOR" or "or exclusive" function. Indeed, this type of scrambling prevents an error from spreading (there is no so-called "avalanche" effect: an error on an element of the scrambled message leads to an error and only one on the message unscrambled) . The avalanche effect is not desirable because it would make it difficult to read the information matrix from the moment when there would be a single error in the scrambled message. As we have seen, the errors play an important role in the implementation of the present invention.

The permutation, step 555, is determined from a permutation algorithm to which a key is provided, said key making it possible to generate pseudo randomly all the permutations carried out. The application of the "exclusive" function, step 560, is between the permuted sequence (whose size corresponds to the number of available cells) and a binary sequence of the same size also generated from a key. Note that if the message is in binary mode (the cells making it possible to represent more than two possible values), the permutation can be done in the same way, and the "exclusive" function can be replaced by a function that modulo adds the number of possible values of the message with a randomly generated pseudo sequence having the same number of possible values as the scrambled message.

Each of the binary data of the scrambled replicated encoded encrypted message is thus modulated in a cell of the information matrix by assigning one of two colors (eg black and white) to the binary data "0" and the other of these colors to a binary "1", the correspondence may vary on the surface of the image.

According to the printing method, step 565, only one of the two colors can be printed, the other corresponding to the original color of the substrate, or having been pre-printed in "backdrop". For printing methods that print a physical relief (for example, embossing or laser engraving), one of the two colors associated with a certain binary value will be chosen, for example arbitrarily.

Alignment blocks, of known or determinable value by the detector, can be inserted in the matrix. These blocks can be inserted at regular intervals from the upper left corner of the matrix, for example every 25 pixels, with a size of 10x10 pixels. It is then observed that the matrix will count 5x5 = 25 alignment blocks, each counting 100 pixels, for a total of 25x100 = 2050 alignment pixels, or 2050 cells of the message. The number of cells available for replication of the encoded message. will then be 12,100-2,500 = 9,600. Given that, as described above, the encoded message has 272 bits, said message can be completely replicated 35 times, and partially a 36 ^th time (the first 80 bits of the encoded message). It is noted that these 35 replications make it possible to improve the signal-to-noise ratio of the encoded signal by more than 15 dB, which allows a very low risk of error when reading the message.

During step 565, a document is marked with the information matrix, for example by printing or etching, with a marking resolution such that the representation of the information matrix includes errors due to said marking step. in such a way that reading of said analog information matrix reveals a non-zero error rate. During this marking step 565, a mark is thus formed comprising, because of the physical marking conditions, local errors, that is to say individually affecting representations of cells of the information matrix, at least partially random or unpredictable.

The physical conditions of marking include, in particular, the physical tolerances of the marking means, support, and, in particular, are surface condition and material, for example ink, optionally deposited. The term unforeseeable means that it can not be determined before the physical marking of the document which cells of the information matrix will be correctly represented by the marking and which cells of the matrix will be erroneous.

For each secret key used, simply change the secret key to renew the initial security level, in case the previous key was discovered by a third party.

It is noted that the encoding and the possible replication make it possible, on the one hand, to significantly increase the robustness of the message in the face of the impairments and, on the other hand, to authenticate the document by estimating or measuring the number or the error rate affecting a reading of the mark of the information matrix.

When we look at captured original information matrices printed at a resolution of 1200 points per inch, with cells of 8 x 8, 4 x 4, 2 x 2 and 1 x 1 pixel (s), we observe that the reading, in high resolution, of the binary value represented by each cell: presents practically no error with cells of 8 x 8 pixels,

- present some errors with cells of 4 x 4 pixels,

presents many errors with cells of 2 x 2 pixels and

for 1 x 1 pixel cells, an error rate approaching the maximum of 50% so that error corrections would probably be insufficient and the degradation due to a copy would be imperceptible since the error rate could not evolve.

Between the extreme dimensions of the cells, one goes through an optimum and, in the limited choice represented here, one of the cases where the cells have 4 x 4 or 2 x 2 pixels is optimal. A method of determining this optimum is given below.

The following is a more detailed description of how optimization of the MIS design is performed according to the printing conditions.

First of all, it is recalled that the IS in digital format, before printing, contains no error. Indeed, there is no random, voluntary or "artificial" generation of errors. These cases are, moreover, not printing errors within the meaning of the present invention.

Thus, it is the printed version of this MIS which contains errors. The errors in question, implemented in the present invention are not artificially induced, but caused in a natural way. Indeed, the errors considered are caused, randomly and naturally, during the marking step, by printing the M IS at a sufficiently high resolution.

These errors are necessary, although their dosage is delicate. Indeed, if the M IS is marked without error (or with a very low error rate), a copy of this MIS made under comparable printing conditions will not involve more error. Thus a printed M IS "almost perfectly" can obviously be copied identically with a means of marking similar. On the other hand, if the MIS is marked with too many errors, only a minority of cells will be copied with additional errors. It is therefore necessary to avoid a marking resolution that is too high, since the possibility of differentiating the originals from the copies is reduced.

Formally, you can not vary the print resolution of an MIS. Indeed, the majority of printing means print in binary (presence or absence of an ink dot) at a fixed resolution, and the gray or color levels are simulated by the different techniques of dithering. In the case of offset printing, this "native" resolution is determined by the resolution of the plate, which is, for example, 2,400 dots / inch (2,400 dpi). Thus, a gray-scale image to be printed at 300 pixels / inch (300 ppi) would actually be printed in binary at 2,400 dpi, each pixel corresponding to approximately 8 x 8 points of the frame.

If we can not generally vary the print resolution, we can, on the other hand, vary the size in pixels of the cells of the M IS, so that a cell is represented by several printing points and , in particular embodiments, the portion of each cell whose appearance is variable, that is to say, printed in black or white, in the binary information matrices. Thus, for example, a cell can be represented by a square block of 1 x 1, 2 x 2, 3 x 3, 4 x 4 or 5 x 5 pixels (non-square blocks are also possible), corresponding to resolutions of, respectively, 2,400, 1,200, 800, 600 and 480 cells / inch.

According to some aspects of the present invention, the number of pixels of the cell leading to a natural degradation in printing is determined to maximize the difference between originals and copies.

Thus, it is preferred to use an error rate between 20 and 25%, since it lies between the copy detection optimums. The optimum of 19.1% corresponds to the case where there is a fixed number of cells, for example if the reading procedure can only read the MIS with a fixed number of cells, whereas the optimum of 27.1% corresponds to the case where there is no constraint on the number of cells, whereas there is a constraint on the physical dimension of the MIS.

It should be noted that, in order to facilitate an autonomous authentication of the information matrix, it is possible to store the decision threshold (s), or other parameters relating to the printing, in the message (s) contained in the information matrix. Thus, it is not necessary to query the database for these parameters, nor to store them on the standalone verification modules. In addition, it makes it possible to manage applications or matrices of information, of the same nature from the application point of view, which are printed by different methods. For example, information matrices applied to the same type of document, but printed on different machines, could use the same key (s). They would have print settings stored in the respective messages.

It is described below how, by measuring the amount of error of the message, a decision on the authenticity of the document can be made based on the amount of error. For this, it is, in principle, necessary to decode the message, because if the message is illegible, we can not determine the errors which it is assigned. However, if the marking has strongly degraded the message (which is particularly the case for copies), or if the amount of information conveyed is high, it is possible that the message is not legible, in which case we can not measure an error rate. It would be desirable to be able to measure the amount of error without having to decode the said message.

On the other hand, the message decoding step implements algorithms that can be expensive. If one does not wish to read the message but only to authenticate, the decoding operation is done only for the purpose of measuring the error rate: it would be preferable to eliminate this step. In addition, if one wishes to make a finer analysis of the error rate, one must reconstruct the replicated message. Replenishing the original replicated message can be expensive and should be avoided.

Now, at the origin of one of the aspects of the present invention it has been discovered that, in order to measure a quantity of errors, it is paradoxically not necessary to reconstruct the original replicated message, nor even to decode the message. It is indeed possible to measure the error quantity of a message by exploiting certain properties of the message itself, at the time of the estimation of the encoded message.

Take the case of a binary message. The encoded message is composed of a series of bits, which are replicated, then scrambled, and the scrambled message is used to constitute the ISM. Interference generally includes a permutation, and optionally the application of an exclusive function, is usually dependent on one or more keys. Thus, each bit of the message can be represented several times in the matrix. During the accumulating step estimation of the encoded message, accumulates the set of indicators of the value of each bit or element of the message. The statistical uncertainty on the value of the bit is generally significantly reduced by this operation. Thus, this estimate, which is considered the correct bit value, can be used to measure the amount of error. Indeed, if the marked matrix has relatively few errors, these will essentially all be corrected during the accumulation step, and thus it is not necessary to reconstruct the encoded message which is already a version without error. In addition, if a few bits of the encoded message have been poorly estimated, generally poorly estimated bits will have a reduced impact on the measurement of the amount of error.

An algorithm of error quantity measurement steps without message decoding is given below for binary data.

- for each bit of the encoded message, accumulate the values of the indicators,

- determining by thresholding the (most probable) value of the bit ("1" or "0"); we obtain the most probable estimate of the encoded message and

- Count the number of indicators (for each cell, it is the density, or normalized luminance value) that correspond to the estimated bit of the corresponding encoded message. It is thus possible to measure an integer number of errors, or a rate or percentage of erroneous bits.

As an alternative to this last step, the value of the indicator can be maintained and a global similarity index can be measured between the values of the indicators and the corresponding estimated bits of the encoded message. A similarity index could be the correlation coefficient, for example.

Alternatively, a weight or coefficient can be associated indicating the probability that each estimated bit of the encoded message is correctly estimated. This weight is used to weight the contributions of each indicator according to the probability that the associated bit is correctly valued. A simple way to implement this approach is to not threshold the accumulations corresponding to each bit of the encoded message.

Note that, the more the message is noisy, the more likely it is that the estimated bit of the encoded message is wrong. This results in a bias such that measuring the amount of error underestimates the amount of actual error. This bias can be estimated statistically and corrected when measuring the amount of error.

It is interesting to note that with this new approach to measuring the amount of error, an MIS can be authenticated without the need to know, directly or indirectly, the messages needed for its design. You just have to know the groups of cells that share common properties.

In variants, several sets of indicators are obtained, from different pretreatments applied to the image (for example, a histogram transformation), or reading at positions different from the M IS; a quantity of errors is calculated for each set of indicators, and the lowest error rate is retained; to speed up the calculations, the encoded message can only be estimated once (this estimate having little chance of changing for each set of indicators).

We can consider that we generate images (or matrices) whose subparts share common properties. In the simplest case, subgroups of cells or pixels have the same value, and they are distributed pseudo-randomly in the image according to a key. The property in question does not need to be known. When reading, we do not need to know this property, since we can estimate it. Thus, the measurement of a score to indicate the authenticity does not require reference to the original image or determination of a message. Thus, in embodiments, the following steps are implemented to perform document authentication:

a step of receiving a set of subgroups of picture elements (for example, pixel values), each subgroup of picture elements sharing the same characteristic, said features not being necessarily known,

an image capture step,

a step of measuring the characteristics of each pixel,

a step of estimating the characteristics common to each subgroup of picture elements,

a step of measurement of correspondence between said estimates of the characteristics common to each subgroup, and said measured characteristics of each of the image elements and

a step of deciding authenticity, as a function of said correspondence measurement.

In other embodiments, which will now be described, to authenticate an MIS, it is not necessary to know or reconstruct the original image, nor to decode the message that it comprises. In fact, it is sufficient, at creation, to create an image composed of subset of pixels that have the same value. At the detection, it is sufficient to know the positions of the pixels that belong to each of the subsets. The property, for example the value of the pixels belonging to the same subset, should not be known: it can be found during playback without the need for message decoding. Even if the property is not found correctly, the M IS can still be authenticated. For the rest, we call this "Random Authentication Pattern"("MAA") this new type of MIS. The word 'random' means that, within a given set of possible values, the MA may take any of its values, without the value being retained after imaging.

For example, suppose that we have a M IS consisting of 12,100 pixels, which is a square of 1 10 x 1 10 pixels. These 12,100 pixels can be divided into 1 1 0 subsets of 1 1 0 pixels each, so that each pixel is in exactly one subset. The division of the pixels into subsets is done in a pseudo-random manner, preferentially using a cryptographic key, so that it is not possible without the key to know the positions of the different pixels belonging to a subset.

Once the 1 10 subsets determined, we assign a random or pseudo-random value to the pixels of each subset. For example, for pixel bit values, the pixels of each subset may be assigned the value "1" or "0" for a total of 1 10 values. In the case of randomly determined values, 1 10 bits are generated with a random generator, these 1 10 bits may or may not be stored thereafter. Note that there are ^2,110 MAAs possible for a given subset division. In the case of pseudo-randomly generated values, a pseudo-random number generator is used which is provided with a cryptographic key, generally stored thereafter. Note that for such a generator based on the hash function SHA1, the key is 160 bits, while we must generate only 1 1 0 bits in our example. Thus, the use of the generator may have limited utility.

Knowing the value of each of the pixels, we can then assemble an image, in our case of 1 10 x 1 10 pixels. The image can be a simple square, increased by a black border facilitating its detection, or can have an arbitrary shape, contain microtext, etc. Groups of pixels with known values for accurate image alignment can also be used.

The image is marked to optimize its degree of degradation, depending on the quality of marking, itself dependent on the quality of the substrate, the accuracy of the marking machine and its settings. Methods are given above for this purpose.

Detection from a captured image of a MAA occurs as follows. Methods of image processing and recognition known to those skilled in the art are applied to accurately locate the pattern in the captured image. Then, the values of each pixel of the MAA are measured (often on a scale of 256 gray levels). For the convenience and consistency of calculations, they can be normalized, for example on a scale of -1 to +1. They are then grouped by corresponding subset, in our example, subsets of 1 1 0 pixels.

Thus, for a subset of pixels having, initially, a given value, there will be 1 10 values. If the value of the original pixels (on a binary scale) was "0", negative values (on a scale of -1 to +1) should dominate, while positive values should dominate if the value was "1". We can then assign the 1 1 0 pixels a value of "1" or "0", and this for each of the 1 10 subassemblies.

For each of the 12,100 pixels, we have a value measured in the image, possibly standardized, and an estimated original value. We can then measure a quantity of error, for example by counting the quantity of pixels that coincide with their estimated value (that is, if the values are normalized on -1 to +1, a negative value, respectively positive, coincides with "0", respectively "1"). We can also measure a correlation index, etc.

The score ("score" meaning both an error rate or a similarity) found is then compared to a threshold to determine whether the captured image corresponds to an original or a copy. Standard statistical methods can be used to determine this threshold.

It is noted that the described procedure does not use external data to the image, in addition to the composition of the subsets, to determine a score. Thus, one can express the count of the amount of error as well.

The quantity of error is equal to the sum, on the subsets, of (Sum (Sign (zij) == f (zi1, .., ziM))).

where z is the (possibly normalized) value of the i-th pixel of the j-th subset with M elements and

f is a function estimating a pixel value for the subset, for example f (z _M , ..., z _iM ) = sign zn-. + ZiM).

In embodiments, such as that shown in Figure 1, the means 105 for obtaining biometric data includes an image sensor 140 for obtaining a photograph of the head of the person or animal and a sensor of image 145 to obtain an image of at least one fingerprint of the person or animal.

The biometric data encoding means 1 is configured to encode the photograph to form the identification image and, optionally, to encode the image of the fingerprint or its characteristic elements, called "minutiae".

The biometric data encoding means 1 is configured to encode the biometric data into at least one hundred thousand elementary cells. Thus, a compressed photograph in four kilobytes can be represented with several redundancies. This photograph can be very precise.

In embodiments, the biometric data encoding means is configured to apply a redundancy factor of at least three and preferably at least five. For example, the identification image illustrated in FIG. 4 comprises sixteen squares with a resolution of 100 per 100 cells that can take one of two shades, ie 160,000 binary elementary data, or bits, which makes it possible to represent 32,000 elementary data bits, or 4,000 bytes, when the redundancy is five.

With such redundancy rates, the detection and correction of printing errors or image capture and image wear or erasure can be achieved even with a lot of lost data. In embodiments, each coding means 10 and 1 is configured to form an image capable of being printed at a resolution of 600 dots per inch while retaining its readability, and the print means 120 prints the dots. identification image and authentication image with this resolution. Thus, the laser printers on the market, which use at least this resolution, can generate secure documents in each company or for each individual.

Preferably, the means 1 1 5 for coding the authentication image is configured to code the image of the fingerprint or its characteristic elements, called "minutiae". Thus, the identity of the holder of the identity document can be verified by comparing each fingerprint data represented by the identification image with the corresponding data of a fingerprint provided during the identification.

In embodiments, the device 10 further comprises a means 150 for forming, on the document, at least one filing surface of a fingerprint. In these embodiments, the capture of at least one fingerprint of the identity document holder is performed directly on the identity document after pressing the finger concerned on the corresponding deposition surface.

In embodiments, the means 1 for coding an authentication image is configured to form a suitable image, upon printing by the printing means, of having at least one tenth of elementary cells whose content is incorrect and, preferably, at least fifteen percent. Thus, the detection of the copy of the document is facilitated because it imposes an error rate of at least fifteen percent (respectively twenty-three percent), and generally at least nineteen percent (respectively twenty seven percent).

In embodiments, the means for encoding an authentication image is configured to encode at least one document holder identification information in the authentication image. Thus, the authentication code, which can not be falsified, identifies the holder of the document, even if the identification image has been falsified.

In embodiments, the means 1 for encoding an authentication image is configured to perform a permutation of elementary cells by implementing a hashing function, or condensate, implementing a hashing key. Thus, the recovery of identification data, by a malicious person, is complicated. Preferably, the hashing key is asymmetrical.

In embodiments, at least one of the means 1 1 0 and 1 15 encoding is configured to perform a permutation of elementary cells representative of a fingerprint or its characteristic elements, by implementing a hash function , or condensate, implementing a hash key. Thus, the recovery of fingerprint data, by a malicious person, is complicated. Moreover, by implementing an inverse function of the hashing function, when reading the identification image thus produced, the error rate is reduced and thus the fingerprint is authenticated.

FIG. 1 also shows a device 20 for verifying the authenticity of an identity document 40 and the identity of a person or an animal. The device 20 comprises: a means 155 for capturing an identification image and an authentication image carried by the document 40, the authentication image comprising random printing errors in the elementary cells of the image of authentication, the identification image representing representative biometric data of that person or animal,

a means 160 for decoding the authentication image,

a means 165 for measuring a rate of erroneous elementary cells in the authentication image,

a means 170 for comparing the rate of erroneous elementary cells with a predetermined limit value,

a means 175 for decoding biometric data encoded in the digital identification image, by implementing redundancies present in the identification image and

a means 180 for displaying the decoded biometric data.

The means 155 for capturing an identification image and an authentication image carried by the document 40 is, for example, consisting of an image sensor with a resolution of at least fourteen megapixels. This sensor can be integrated in a specific reader or in a smartphone.

The means 1 60 for decoding the authentication image is adapted to decode the authentication data, by implementing the redundancies.

The means 165 for measuring a rate of erroneous elementary cells in the authentication image can compare the captured image and a reconstructed image by recoding the authentication data or can count the number of redundancies used to decode the data of the authentication data. 'authentication. This number of erroneous cells compared to the total number of cells gives a rate of erroneous elementary cells.

The comparison means compares the rate of erroneous elementary cells with a predetermined limit value. If the erroneous cell count is greater than the predetermined limit value (for example 15 or 23 percent, as discussed above for print error rates of 10 and 15 percent), document 40 is considered as falsified.

The means 1 for decoding biometric data coded in the digital identification image, by implementing redundancies present in the identification image, provides the digital biometric data (size, color of the hair and eyes, for example ) and image data (face and / or fingerprint, for example).

The means 180 for displaying the decoded biometric data is, for example consisting of a specific reader screen, computer or smartphone.

Preferably, the device 20 comprises a central unit 185 for comparing the identification biometric data with data captured on the document holder 40.

In particular, by extracting the characteristic elements of the document holder's fingerprints 40, captured on the fingerprint deposition surface, and then comparing these extracted elements with the fingerprint characteristic elements coded in at least one of the two identification and authentication images, the central unit 185 determines whether the bearer of the document 40 is the carrier that this document 40 identifies. FIG. 2 shows the method of producing a document supporting an identification of a person or an animal. This method comprises:

a step 205 for obtaining biometric data representative of this person or this animal,

a step 220 of encoding the biometric data to form a digital identification image comprising redundancies,

a coding step 225 for forming an authentication image of the document, said authentication image presenting elementary cells and

a step 230 of printing the authentication image and the identification image on the document to generate random errors in the printed elementary cells of the authentication image.

Step 205 comprises a step 21 0 for capturing biometric data by implementing at least one sensor and a step 215 for receiving biometric data stored in memories.

Step 220 encodes the biometric data with redundancies, for example error detection and correction codes ("CRC" for "check redundancy code").

Step 225 codes identification data preferentially with redundancies, for example error detection and correction codes, CRC.

FIG. 3 shows the method 50 of verifying the authenticity of an identity and identity document of a person or an animal. The method 50 comprises:

a step 305 for capturing an identification image and an authentication image carried by the document, the authentication image comprising random printing errors in the elementary cells of the authentication image the identification image representing representative biometric data of that person or animal,

a step 310 of decoding the authentication image,

a step 315 of measuring a rate of erroneous elementary cells in the authentication image,

a step 320 for comparing the rate of erroneous elementary cells with a predetermined limit value,

a step 325 of decoding biometric data encoded in the digital identification image, by implementing redundancies present in the identification image,

a step 330 of biometric data capture of the document holder 40,

a step 335 of checking correspondence between the captured biometric data and the decoded biometric data and

a step 340 of displaying the decoded biometric data.

Thus, the malicious people can not have access to the data represented by the identification image. The falsification of the document is thus made more difficult. In addition, only the holder of the document can thus have access to data stored in the document.

For the implementation of the functions of hashing, the skilled person can refer to the article "Symmetric Hash Functions for Fingerprint Minutiae" by Sergey Tulyakov, Faisal Farooq and Came Govindaraju, SUNY at Buffalo 14228, New York State, United States of America, and the bibliographic references cited therein.

FIG. 4 shows a document 40 for identifying a person or an animal, which comprises:

a digital identification image 405 comprising redundancies and representative of biometric data representative of a person or animal,

an authentication image 410 presenting elementary cells and random printing errors in the elementary cells and

a surface 415 of fingerprint deposition.

The deposition surface 41 is, for example, made up of two surfaces located near the edge of the document 40 so that the wearer applies his thumbs to this surface 415 when he delivers the document 40 to a person responsible for its verification.

The surface 415 can be cleaned by simply passing a tissue or tissue. For example, surface 41 is a metallized or black smooth surface.

Each code preferably includes locating elements 420 to make it possible to scale and rotate the captured image in order to decode it by identifying each cell.

In the embodiment illustrated in FIG. 4, the document 40 is a microcircuit card.

425.

As can be understood from the above description, the present invention can be implemented for very secure official documents but also for private documents generated and printed by means available to the general public, such as personal computers. and laser printers.

In addition, the verification of authenticity and the obtaining of the biometric data can be performed on a non-dedicated terminal, for example with a computer application downloaded to a mobile phone provided with an image sensor of sufficient resolution.

It is noted that the capture of the authentication image, of higher resolution, can be performed using the same image sensor as the capture of the identification image, but at a distance from the weaker document.

Claims

1. Device (10) for producing a document (40) supporting an identification of a person or an animal, characterized in that it comprises:

means (1 90) for determining printing conditions of said document;

means (105) for obtaining biometric data representative of this person or animal,

means (1 10) for coding the biometric data to form a digital identification image comprising redundancies,

means (195) for determining the physical characteristics of cells of at least one shape, as a function of the printing conditions, in such a way that the proportion of cells printed with a printing error coming exclusively from the printing hazards is greater than a first predetermined value and less than a second predetermined value;

coding means (1 15) for forming, in said form, an authentication image of the document, said authentication image having elementary cells whose content is representative of the coded authentication image and

means (120) for printing the authentication image and the identification image on the document, by implementing said printing conditions, to generate random errors in the printed elementary cells of the document; authentication image said form being adapted to allow the detection of a copy modifying the appearance of a plurality of said elementary cells.

2. Device (10) according to claim 1, wherein the means (1 05) for obtaining biometric data is configured to obtain a photograph of the head of the person or animal and the coding means of the biometric data. is configured to encode the photograph to form the identification image.

3. Device (10) according to one of claims 1 or 2, wherein the means (1 10) for encoding biometric data is configured to encode the biometric data in at least one hundred thousand elementary cells.

4. Device (10) according to one of claims 1 to 3, wherein the means (1 10) for encoding biometric data is configured to apply a redundancy factor of at least three.

The device (10) according to one of claims 1 to 4, wherein each coding means (1 10, 1 1 5) is configured to form a printable image with a resolution of 600 dots per inch. retaining his reading skills.

6. Device (10) according to one of claims 1 to 5, wherein the means (105) for obtaining biometric data is configured to obtain data representative of at least one fingerprint of the person and the means ( 1 15) biometric data encoding is configured to encode the data representative of the fingerprint.

7. Device (10) according to one of claims 1 to 6, which comprises means (1 50) forming on the document of at least one fingerprint deposition surface.

8. Device (10) according to one of claims 1 to 7, wherein the means (1 15) for encoding an authentication image is configured to form a suitable image, when printing by the printing means (120), comprising at least one tenth of elementary cells whose content is erroneous.

9. Device (10) according to one of claims 1 to 8, wherein the means (1 15) for encoding an authentication image is configured to encode at least one identification of the document holder in the document. authentication image.

10. Device (10) according to one of claims 1 to 9, wherein the means (1 15) for encoding an authentication image is configured to perform a permutation of elementary cells by implementing a hash function , or condensate, implementing a hash key.

1 1. Device (10) according to one of claims 1 to 10, wherein at least one of the means (1 1 0, 1 15) encoding is configured to perform a permutation of elementary cells representative of a fingerprint by putting implement a hash function, or condensate, implementing a hash key.

12. Device (20) for verifying the authenticity of a document of identity and identity of a person or an animal, characterized in that it comprises:

means (1 55) for capturing an identification image and an authentication image carried by the document, the authentication image comprising elementary cells and random printing errors in the elementary cells , the proportion of cells printed with a printing error coming exclusively from the print randomness is greater than a first predetermined value and less than a second predetermined value, the identification image representing biometric data representative of this person or of this animal,

means (160) for decoding the authentication image,

means (165) for measuring a rate of erroneous elementary cells in the authentication image,

means (1 70) for comparing the rate of erroneous elementary cells with a predetermined limit value,

means (175) for decoding biometric data encoded in the digital identification image, by implementing redundancies present in the identification image and

means (180) for displaying the decoded biometric data.

13. A method (30) for producing a document supporting an identification of a person or an animal, characterized in that it comprises:

a step of determining conditions for printing said document;

a step (205) for obtaining biometric data representative of this person or this animal,

a step (220) for coding the biometric data to form a digital identification image comprising redundancies,

a step of determining the physical characteristics of cells of at least one shape, as a function of the printing conditions, in such a way that the proportion of cells printed with a printing error coming exclusively from the printing hazards is greater than a first predetermined value and less than a second predetermined value; a coding step (225) for forming, in said form, an authentication image of the document, said authentication image having elementary cells whose content is representative of the coded authentication image and

a step (230) for printing the authentication image and the identification image on the document, by implementing said printing conditions, to generate random errors in the printed elementary cells of the document; authentication image said form being adapted to allow the detection of a copy modifying the appearance of a plurality of said elementary cells.

14. Method (50) for verifying the authenticity of an identity and identity document of a person or an animal, characterized in that it comprises:

a step (305) for capturing an identification image and an authentication image carried by the document, the authentication image comprising elementary cells and random printing errors in the elementary cells, the proportion of cells printed with a printing error originating exclusively from the print randomness is greater than a first predetermined value and less than a second predetermined value, the identification image representing biometric data representative of this person or of this animal,

a step (31 0) of decoding the authentication image,

a step (315) of measuring a rate of erroneous elementary cells in the authentication image,

a step (320) of comparing the rate of erroneous elementary cells with a predetermined limit value,

a step (325) for decoding biometric data encoded in the digital identification image, by implementing redundancies present in the identification image and

a step (340) for displaying the decoded biometric data.

15. The method of claim 14, which comprises the step (340) of displaying the biometric data after a step (330) for capturing biometric data of the document holder and a step (335) for verifying correspondence between the data. captured biometric and decoded biometric data.

16. Document (40) for the identification of a person or animal that includes:

a digital identification image (405) comprising redundancies and representative of biometric data representative of a person or animal,

an authentication image (410) presenting elementary cells and random printing errors in the elementary cells, the proportion of cells printed with a printing error originating exclusively from the printing errors being greater than a first predetermined value; and less than a second predetermined value.

17. Document (40) according to claim 16, which comprises a surface (415) depositing at least one fingerprint.