• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to a method for matching (100) representations of sets of planar related graphs (110), each graph (110) of a set comprising at least two vertices and an edge connecting the vertices, the method comprising for each set of graphs, steps of: - determining (200) a structure of at least one graph (110) of the set as a function of the number of edges to which the vertices are related, - definition (300) ) a so-called topographic code of the determined structure, - comparison (400) of the topographic code defined with a resulting predetermined topographic code, resulting in a qualification of the correspondence, according to predetermined thresholds, in low (rejection), acceptable (authentication ) or strong (rejection). The invention also relates to a computer program for executing the method, a system for implementing the method as well as the use of the method in the field of biometric authentication and electronic mapping.
    公开号:CH713468A2
    申请号:CH00120/18
    申请日:2018-02-02
    公开日:2018-08-15
    发明作者:Rossi Mathieu
    申请人:Rossi Mathieu；
    IPC主号:
    专利说明:

    Description
    TECHNICAL FIELD OF THE INVENTION
    The present invention relates to a mapping process of representations of sets of planar related graphs, each set of graphs may differ from another set of graphs by the number and structure of the graphs it contains. More particularly, the invention relates, by its fields of application, the field of biometrics (authentication and / or identification) when these graphs are derived from skeletonized images of venous networks or else the field of electronic mapping (implementation cards) when these graphs come from skeletonized aerial shots. The invention also relates to a system for implementing the mapping method.
    STATE OF THE PRIOR ART
    [0002] Authentication techniques using palm veins have been the subject of studies and research for almost twenty years. Indeed the palm of the hand, like the periorbital area of the face and the arch, has a large and complex vascularization, allowing a precise identification. Moreover, thanks to the absence of hairs of the palm of the hand and the arch, the collection of the data relating to the venous network which they cover is easy. In addition, this collection is hygienic because shooting is possible without contact with the skin. Finally, because of their subcutaneous implantation, the veins are very difficult to counterfeit, unlike fingerprints.
    [0003] The prior art discloses techniques for implementing this type of authentication; these are based on the mapping of venous network arrangements and described below: - Venous network photography: the deoxidized hemoglobin circulating in the veins absorbs light from the near-infrared zone; a photograph using this type of lighting produces black lines at the location of the veins. - The extraction of the area of interest (ROI = Region of Interest in English): the recognition and the isolation of this zone, generally rectangular, is carried out starting from the examination of the contour of the hand and makes it possible to limit the effects of rotation and translation induced during shooting. - Pretreatment of the image: its objective is to reveal on a two-color image only the venous network. This is achieved by applying a succession of appropriate filters using computer programs of image processing. - The extraction of feature points (feature extraction in English): its purpose is to digitize characteristics of the venous network to keep them in databases for authentication operations, ie for the purpose of comparing a new shot with those present in the base. SIFT (Scale-Invariant Feature Transform in English), Hessian matrices and transformed into wavelets or curvelets are among the techniques of extraction of characteristic points often cited. - Enlistment: it consists of storing in a database the extracted characteristic points. - Authentication: it consists in acquiring an image of a venous network, in processing the above mentioned processing steps and in searching the database if the acquired image corresponds to one of the already recorded images. The comparison techniques used in this authentication step are closely related to the techniques used for the extraction of characteristic points.
    The common point of most of the methods taught by the literature is to use image processing techniques to extract from said images many textural characteristics, techniques requiring for the most part a phase of extraction of the d-zone. interest (ROI), which eliminates the effects of rotation and translation. All these techniques, derived from pattern recognition research, are therefore interested in the texture of the venous network.
    [0005] Some documents propose, for shape recognition problems, methods of extracting characteristic points in which graphs are constructed: the vertices then represent the texture elements of the image ("regions"), the edges represent the relations between these regions and - the contiguous labels provide information on the position of the regions, gradient orientation histograms or texture elements.
    These extraction methods lead to use, for comparisons, methods of matching graphs. The literature proposes different approaches but the associated processing times vary exponentially depending on the number of inputs and the proposed algorithm tracks require power and processing times that are incompatible with real-time use.
    Whatever the approach considered for the extraction of the characteristic points, texture elements of the image are used, thus making the associated processing methods dependent on maximum tolerance to a rotation between the acquired image. and the reference image and, therefore, binding for the shooting conditions.
    There is therefore a real need to create an authentication method and / or identification optimized at the same time processing time, simplicity of shooting and level of discrimination.
    SUMMARY OF THE INVENTION
    The objective of the present invention is therefore to propose, by focusing on the structure of the venous networks rather than on their texture and by comparing binary codes derived from these structures, a faster authentication method and more efficient than those of the state of the art.
    The invention thus relates, according to a first aspect, to a method for mapping (100) representations of sets of planar related graphs (110), each of the graphs (110) of a set comprising minus two vertices (T, C) and an edge linking said vertices, the method comprising, for each set, the following steps: • determination (200) of a structure (270) of at least one graph (110) of the set according to the number of edges to which the vertices are related, • definition (300) of a so-called topographic code (360) of the determined structure (270) comprising the following steps: - assembly (330) of the coding elements Identical principals (ECPI), - Cardinal determination (350) (NOcc) of each of the identical primary coding element assemblies (ECPI), • Comparison (400) of the defined topographic code (360) with a pre-established topographic code (390) ), said comparison (400) comprises the next steps: - determining (410) the presence rates (TP) within the pre-established topographic code (390) of at least a portion of the identical primary coding element (ECPI) assemblies of the defined topographic code (360); ); developing (420) a measure of similarity between the defined topographic code (360) and the preset topographic code (390) by a function (Sim (TP, NOcc)), involving the presence rates (TP) of identical primary coding elements (ECPI) and the cardinal (NOcc) of each of them; - Qualifying (430) the comparison (400) in weak, acceptable or strong correspondence according to the value of the similarity measure with respect to predetermined thresholds.
    According to particular features, the step of determining the structure of the at least one graph of the set comprises a step of characterizing its vertices in: - terminal vertices, for the vertices linked to the graph by a single edge, and - vertexes of crossing, for the vertices connected to the graph by several edges.
    According to particular features, the determination of the structure of the at least one graph further comprises a step of identifying at least a portion of the different oriented paths: - existing within the at least one graph, the paths beginning with a vertex called of origin and ending with a summit called of destination, - and within which each edge is traversed only once.
    According to particular features, the determination of the structure of the at least one graph further comprises, when a registered oriented path comprises at least one intermediate crossing point which is neither the original vertex nor the destination vertex, a calculation step, for the at least one intermediate crossing vertex: the number of edges coming from the at least one intermediate crossing vertex, not belonging to the oriented path and located between the ridge of the path arriving at the at least one intermediate crossing vertex and the ridge of the path extending from the at least one intermediate crossing vertex by scanning the graphical representation counter-clockwise from the ridge arriving at the at least one intermediate crossing vertex, said stops being numbered from 1 in the scanning direction, and - the number of ridges from the at least one cross vertex intermediate level, not belonging to the oriented path and being situated between the ridge of the road arriving at the at least one intermediate crossing point and the ridge of the path leading from the at least one intermediate crossing vertex by sweeping the graphical representation in the clockwise direction from the ridge arriving at the at least one intermediate crossing vertex, said edges being numbered from 1 in the scanning direction.
    According to particular features, the step of defining the topographic code of the determined structure comprises a step of assigning a principal coding element to each identified oriented path according to the main characteristics of the identified oriented path, the main characteristics being: - the nature of the vertices and / or - the number and the situation of the edges coming from the at least one intermediate crossing vertex.
    According to particular features of the invention, when the representations of sets of graphs are derived from skeletonized images, the definition of the topographic code of the determined structure further comprises a step of assigning a coding element said appendix to each indexed path listed according to ancillary features of the identified directed path, the ancillary features being: - the ratio between the number of pixels of an edge of the listed directed path appearing on the graphical representation and the number of pixels joining in the two vertices of the said ridge, - for each intermediate junction of the marked lane, the indication of the angle formed by the direction of the ridge coming from the intermediate junction with respect to the direction of the ridge arriving at that junction. intermediate crossing, - for each intermediate crossing point of the identified marked path, where the number of edges resulting from this intermediate crossing vertex is non-zero, the ratio existing between, on the one hand, the value of the angles formed by the first edge encountered in each of the scanning directions with the edge arriving at the intermediate crossing vertex and, on the other hand, the value of the angles formed by the last edge encountered in each of the scanning directions with the edge coming from the intermediate crossing vertex.
    According to particular features, when the representation of the graphs originates from a skeletonized image, the comparison of the topographic code defined with the pre-established topographic code furthermore comprises a consideration, in the development of the similarity measure, additional coding elements.
    The present invention relates, in a second aspect, to a computer program product comprising code instructions for executing a matching method according to the first aspect when this program is executed by a processor. .
    The present invention relates, according to a third aspect, to a system for identifying and / or authenticating an individual for implementing the method according to the first aspect, comprising: a production device consisting of a capture device dedicated to the acquisition of digital images of biometric data in the visible domain and / or in the infrared domain and a data processing device configured to process the digital image acquired by the capture device, being able to extract a skeletonized image from the digital image, characterized in that it is furthermore capable of generating a set of planar graphs from the skeletonized image and to be executed on this set of graphs the steps of determining and defining the method according to the first aspect, and - an evaluation device, consisting of a storage device storing at least one pre-established topographic code and a comparison device, characterized in that it makes it possible to implement the steps of the method according to the first aspect of developing and qualifying a measure of similarity between a defined binary code provided by the production device and a pre-established binary code present in the storage device.
    According to particular features: - the data processing device comprises a computer and communication interfaces with the capture and evaluation devices, - the storage device is made using any type memory - RAM, Flash, hard disk or SSD, for example - and the binary codes are organized into databases. the capture, data processing, storage and comparison devices comprise computers and communication interfaces connecting them to each other.
    According to particular characteristics: the computers consist of processors or microprocessors, presented for example of the RISC or x86 type, presented for example in the form of controllers, microcontrollers, DSPs, FPGAs or any other combination allowing to implement the steps of carrying out the method according to the invention and, where appropriate, adding to the topographic bit code all relevant context information for the implementation in question, in particular the precise geographical position (GPS or equivalent) and the date and time of the digital image capture, - the communication interfaces between devices are wired or wireless and use communication protocols such as RS-232, USB, FireWire, HDMI, Bluetooth, ZigBee, Ethernet, Wifi or mobile telephony , satellite or fixed.
    According to particular features, the capture and data processing devices can be integrated within the same housing, such as a mobile phone.
    The invention relates, according to a fourth aspect, to the use of the mapping method of representations of sets of planar related graphs according to the first aspect, in the field of biometric authentication, when these graphs are derived from skeletonized images of venous networks and / or in the field of electronic mapping when these graphs are derived from skeletonized aerial photographs.
    BRIEF DESCRIPTION OF THE FIGURES
    The invention will be better understood on reading the description which follows and the examination of the figures that accompany it. These are presented for illustrative purposes only but not limited to the invention. The figures show: fig. 1 a representation, in the form of a logic diagram, of an embodiment of the method according to the invention; Fig. 2 a representation, in the form of a logic diagram, of a step of determining the same embodiment of the method according to the invention; Fig. 3 a representation, in the form of a logic diagram, of a step of defining the same embodiment of the method according to the invention; Fig. 4 a representation, in the form of a logic diagram, of a comparison step of the same embodiment of the method according to the invention; Fig. A diagrammatic representation of material means of an embodiment of the method according to the invention.
    In the description, the same reference numerals designate from one figure to the other identical or functionally similar elements.
    DETAILED DESCRIPTION OF THE EMBODIMENTS OF THE INVENTION
    FIG. 1 shows, in the form of a logic diagram, an embodiment of the method 100 for mapping representations of sets of planar related graphs according to the invention. In the example, the representation of sets of graphs comes from skeletonized images and each graph of a set comprises at least two vertices and an edge linking the vertices.
    Skeleton pretreatment aims to transform a captured image into a two-color image sque-lettisée. This pretreatment may for example be carried out using standard image processing methods, successively applying standardization processes and threshold and contour filters. Once the skeletonized image is obtained, the extraction processing of the planar related graphs of this image is obtained by the application of classical algorithms.
    In one embodiment, one may for example: • associate an orthonormal frame (O, x, y) with the skeletonized image, the unit on the abscissa as on the ordinate being the pixel, in which the position of each pixel P of the skeleton is determined by its coordinates (x, y) in this frame; Defining within the set of pixels of the image a neighborhood relation V such that two pixels Pi (xi, yi) and P2 (x2, y2) will be declared neighbors under this relation if and only if lx2- XiI <1 and I y2-y21 <1; • define the notion of "neighborhood chain": a set of pixels P ,, with ie [1, n], form a neighborhood chain C, of extremities [Pi, Pn], if and only if Vi e [1 , n-1], Pi = V (Pi + i); Partitioning the pixels belonging to the skeletonized image representing a venous network into a set R of disjoint related networks R, such that for any pair of pixels (Pk, Pi) e R, there exists a neighborhood chain connecting Pk to Pi ; Associating with each connected network R a graph G 1 (V 1, E 2) whose vertices v e V are the pixels having either a single neighboring pixel or more than 2 neighboring pixels; and whose edges e e E, are the neighborhood chains joining the vertex pixels. We will call G the set of graphs thus associated with networks R, e R.
    The method 100 which is the subject of the present invention comprises, for a graph 110, steps of: determination of its structure 270 as a function of the number, here twelve, of edges to which its vertices are linked, here eleven; definition 300 of a topographic code of the determined structure 270; - comparison 400 of the topographic code defined with a pre-established topographic code.
    Fig. 2 details this determination 200, which comprises a step of characterizing 210 of its eleven vertices in: six terminal vertices T, for vertices linked to the graph by a single edge, and five crossing vertices C, for those linked to the graph by several edges.
    In one embodiment, it is possible, for example, for each of these graphs Gi (Vj, Ej), to characterize the vertices veVj in vertex ends T and of intersection C, for example by determining the two subsets ViT and ViCde Vj such that: 0 vT s ViT <=> (vT s K) a (3! Vg V "(vT, v) e £.) 0 vcG ^ cC -> (vceK) A (vcîZ- ^ r) [ 0031] The determination 200 further comprises a step 230 of identifying at least a portion of the different oriented paths existing within the graph 110: the paths starting with a vertex T or C said of origin, and ending with a vertex T or C said of destination, - and in which each edge is traversed only once.
    In one embodiment, it is possible, for example, to identify the sets HIT and HIC of paths oriented h = (v0, ei, Vi ... en-i, Vn-i, en, Vn) of V, all of which vertices are distinct (except possibly v0 and vn), all edges of which are distinct, such as: 0 hj. e HiT "(v0 g ViT) A (v" g VjT) λ (vkM1 n_l} g ViC) a (n> 2) λ (v0 * v ") o hc G HiC O (v0 G riC) A (y" The determination 200 of the structure 270 of the graph 110 further comprises, when a listed directed path CH has at least one, here two, top (s) of intersection (s) said intermediate (s) Cl, which is neither the original vertex nor the destination top, a calculation step 250, for each intermediate crossing vertex Cl: - of the number nD of edges coming from the intermediate crossing vertex Cl not belonging to the path oriented CH and being situated between the edge of the path arriving at the intermediate crossing vertex C1 and the edge of the path starting from the vertex of intermediate crossing C1 by scanning the graphical representation counter-clockwise from the edge arriving at the intermediate crossing vertex C1, said edges being numbered from 1 in the direction of the sweep (numbers from 1 to nD), and - the number nG of edges from the intermediate crossing vertex C1 that do not belong to the path CH and located between the edge of the path arriving at the intermediate crossing vertex Cl and the edge of the path from the intermediate crossing vertex C1 by scanning the graphical representation clockwise from the ridge arriving at the intermediate crossing vertex C1, said edges being numbered from 1 in the scanning direction (numbers from 1 to nG).
    In one embodiment, it will be possible to calculate for each intermediate vertex vk, ke [1, n-1], the following elements: nkD: number of incident edges in vk, distinct from ek and ek + 1 , located to the right of the road; - nkG: number of incident edges in vk, distinct from ek and ek + 1, located to the left of the path; pk: ratio of "curvature" of ek (ratio between the number of pixels of the skeleton of ek and the number of pixels necessary to join in a straight line vk_i to vk; - yk: angle (ek, ek + 1) - akiG, i <i <nG: angles formed with ek by the incident edges to the left of the path; - akiD, i <i <nD: angles formed with ek by incident edges to the right of the path.
    FIG. 3 details the definition 300 of a topographic code 360 of the determined structure 270. Said definition 300 of the topographic code 360 comprises an allocation step 310 of a coding element called ECP principal to the identified directed path CH according to its characteristics. main 260, namely the nature of the vertices T and C and / or the numbers nG and nD and the situation of the edges from each intermediate crossing vertex C1.
    The definition step 300 of the topographic code 360 further comprises assembly steps 330 of the identical primary coding elements ECPI and determination 350 of the cardinal NOcc of each of the identical primary coding element assemblies ECPI.
    The definition 300 of the topographic code 360 furthermore comprises a step 315 of allocating a coding element called ECA appendix to the identified directed path CH according to its associated characteristics 265, namely: the ratio p between the number of pixels of an edge of the CH oriented path and the number of pixels joining in a straight line the two vertices T and / or C of said edge, - for each intermediate crossing C1 of the listed directed path CH, the indication y of the An angle formed by the direction of the edge of the intermediate intersection C1 with respect to the direction of the ridge arriving at this intermediate intersection C1, for each intermediate crossing point C1 of the identified directed path CH, when the number nG and / or nD edges from this intermediate crossing vertex C1 is non-zero, the value of the angles a1G and a1D formed by the first edge encountered in each of the scanning directions with the edge arriving at intermediate crossing vertex C1 and the value of the angles αΝα and aND formed by the last edge encountered in each of the scanning directions with the edge coming from the intermediate crossing vertex C1.
    In one embodiment, it will be possible to carry out the operations described hereinafter. • Assign a main encoding element to each path in the form ^ C / 'jJ. = (7 /;%,% e (riQ,! / Ty |,' iG, niD) '2G,' 2G, '2Z) , ..., "(" _ 1) G, "(,, _ i) D); ECP, the set of the main coding elements (ECP) of the graph G, and ECP will be called the set of the main coding elements of the set G of graphs considered. • Group the main coding elements (ECP) into classes (ECPI), two ECP ECP, p and ECP.q belonging to the same class if all their coordinates are equal one by one. The vector
    Can represent their common class, its first element being the number of identical ECPs grouped in the class and its other elements being the values common to all vectors of the class. ECPI is the set of identical main coding elements of the set G of graphs considered. • Assign an additional code element to each path h, j & HiT <jHiC, in the form ^ C4.y = (Æ ^ Xe (rC), | Ai7. |, Uls4s "_i (pt, / t, ixh. G ^ o)). ECA will be called the set of subsidiary coding elements of the set G of graphs considered. • The topographic code of the set G of graphs is the union CT - ECPI <JECA.
    FIG. 4 details the comparison 400 of the topographic code defined 360 with the preset topographic code 390. Said comparison 400 comprises steps of: - determination 410 of the presence rates TP, within the pre-established topographic code 390, of at least a part of the assemblies identical primary coding elements ECPI of the topographic code defined 360; developing a measure of similarity between the defined topographic code 360 and the pre-established topographic code 390 by a function Sim (TP, NOcc), involving the presence ratios TP of identical main coding elements ECPI and the cardinal NOcc of each of them; - Qualification 430 of comparison 400 in low correspondence, acceptable or strong depending on the value of a similarity measure against predetermined thresholds.
    The comparison step 400 of the topographic code defined 360 with the pre-established topographic code 390 further includes a consideration, in the development 420 of the similarity measurement, ECA ancillary coding elements.
    In one embodiment, it is possible, for example, to perform the operations described below: • Name ECPID the topographic code defined 360 and ECPIP the preset topographic code 390. • Sort the set ECPID by decreasing values of the number NOcc d occurrences. • Establish the TP attendance rate of all ECPId elements within ECPIP. This rate is, for each element of ECPId, the minimum of the NOcc of the element of ECPId and the NOcc of the corresponding element in ECPIP. This number can be zero if the ECPID element is not present in ECPIP. • Establish the measure of similarity between the candidate code and each pre-established code. This measurement is the ratio between the sum of the TP and the sum of the NOcc of the ECPID code. • Complete the similarity measure by adding the rate (weighted by a predetermined K coefficient) of the presence of the coding elements of the ECPId code within ECPIP. • Compare the similarity measure thus completed with 2 predetermined SIM2 and SIM3 thresholds and generate the following answer: - Low match: the similarity measure is less than SIM2 - Acceptable match: the similarity measure is between SIM2 and SIM3
    权利要求:
    Claims (13)
    [1]
    - Strong match: the similarity measure is greater than SIM3. In one embodiment, schematized in FIG. 5, the system 500 according to the invention comprises in particular a production device 510 and an evaluation device 550. The production device 510 notably comprises a capture device 520, dedicated to the acquisition of useful biometric data 525 for the implementation of the method according to the invention. Such a capture device may especially be in the form of a camera or a camera for capturing digital images in the visible range and / or in the infrared range. The production device 510 further comprises a data processing device 530 configured to process the image of the venous network acquired by the capture device 520. Such processing notably comprises a so-called skeletonization pretreatment of the captured image. , extracting from the skeletonized image 533 of a set of related planar graphs, determining the structure of these graphs and defining a binary code 537 representing the structure. The data processing device 530 may comprise a computer and communication interfaces with the capture devices 520 and evaluation 550. The evaluation device 550 comprises in particular a storage device 560, configured to store at least a binary code from a reference vein network. In one embodiment, this storage device 560 can be implemented using any type of memory - RAM, Flash, hard disk or SSD, for example - and the binary codes can be organized into databases . The evaluation device 550 further comprises a comparison device 570 for developing and qualifying a measure of similarity between a binary code provided by the production device 510 and a binary code present in the storage device 560. In the device according to the invention, the capture devices 520, data processing 530, storage 560 and comparison 570 can comprise computers and communication interfaces connecting them to each other. The computers may consist of processors or microprocessors, for example of the RISC or x86 type, presented for example in the form of a controller, microcontroller, DSP, FPGA or any other combination making it possible to implement the steps of the processes described herein. before and, if necessary, add to the topographic bit code all relevant context information for the implementation considered, including the precise geographical position (GPS or equivalent) and the date and time of the image capture. The communication interfaces between devices may for example be of the wired or wireless type and use communication protocols such as RS-232, USB, FireWire, HDMI, Bluetooth, ZigBee, Ethernet, Wifi or mobile telephony, satellite or fixed. In one embodiment of the device according to the invention, the capture devices 520 and data processing 530 can be integrated within the same housing, for example a mobile phone. One of the advantages of the method according to the present invention is that it is asymmetrical, thus preventing any reconstruction of the image from the coding. Indeed, it is the code and not the image that will be transported to a database, whether at the time of the initial registration or during an authentication request, the image itself never leaves the local device of shooting (or sensor), in which it is destroyed immediately after generation of the code. The method according to the present invention is tolerant to image differences between registration and authentication, differences that can be caused by accidental alteration of a venous network - for example as a result of an injury - or a difference in the lighting conditions of the shooting, or a different positioning of the biometric fingerprint with respect to the sensor (rotation relative to the axis of the camera, different angle with the plane perpendicular to the axis of the camera). camera), or other situations. In addition, the method, object of the present invention is highly discriminating and generates minimum rates of false positives and false negatives. In addition, it is efficient and effective, because it allows to quickly produce the results of an authentication, even when the authentication is performed vis-à-vis a database with several million registered codes. claims
    A method of mapping (100) representations of sets of planar related graphs (110), each of graphs (110) of a set comprising at least two vertices (T, C) and an edge linking said vertices, the method comprising, for each set, the following steps: • determination (200) of a structure (270) of at least one graph (110) of the set as a function of the number of edges to which the vertices are linked, Defining (300) a so-called topographic code (360) of the determined structure (270) comprising the following steps: - assembly (330) of the identical principal coding elements (ECPI), - determination (350) of the cardinal (NOcc) ) of each of the identical primary coding element assemblies (ECPI), • comparing (400) the defined topographic code (360) with a pre-established topographic code (390), said comparing (400) comprises the following steps: - determining ( 410) attendance rates (TP) within the pre-established topographic code (390), at least a portion of the identical primary coding element (ECPI) assemblies of the defined topographic code (360); - developing (420) a measure of similarity between the defined topographic code (360) and the pre-established topographic code (390) by a function (Sim (TP, NOcc)); involving the presence rates (TP) of identical main coding elements (ECPI) and the cardinal (NOcc) of each of them; - Qualifying (430) the comparison (400) in weak, acceptable or strong correspondence according to the value of the similarity measure with respect to predetermined thresholds.
    [2]
    Method (100) according to claim 1, wherein the determination (200) of the structure (270) of the at least one graph (110) of the set comprises a step of characterizing (210) its vertices ( T, C) in: - terminal vertices (T), for vertices linked to the graph by a single edge, and - vertex crossings (C), for vertices linked to the graph by several edges.
    [3]
    The method (100) of claim 1 or 2, wherein the determining (200) of the structure (270) of the at least one graph (110) further comprises a step (230) of enumerating (230) a part of the different oriented paths (CH) existing within the at least one graph (110): the paths beginning with a vertex (T, C) said of origin and ending with a vertex (T, C) - destination, 0 and in which each edge is traversed only once.
    [4]
    The method (100) of claim 3, wherein the determining (200) of the structure (270) of the at least one graph (110) further comprises, when a referenced directed path (CH) comprises at least an intermediate said crossing vertex (Cl) which is neither the original vertex nor the destination vertex, a calculation step (250), for the at least one intermediate crossing vertex (Cl): - of the number nD edges from the at least one intermediate crossing vertex (Cl), not belonging to the oriented path (CH) and being located between the edge of the path arriving at the at least one intermediate crossing vertex ( Cl) and the edge of the path starting from the at least one intermediate crossing vertex (Cl) by scanning the graphical representation counterclockwise from the edge arriving at the at least one intermediate crossing vertex (Cl) ), said edges being numbered from 1 in the sense of the balaya ge, numbers from 1 to nD, and - the number nG of edges coming from the at least one intermediate crossing vertex (Cl), not belonging to the oriented path (CH) and being situated between the edge of the path arriving at the at least one intermediate crossing vertex (Cl) and the road edge from the at least one intermediate crossing vertex (Cl) by scanning the graphical representation clockwise from the arriving ridge at the at least one intermediate crossing vertex (Cl), said edges being numbered from 1 in the scanning direction, numbers from 1 to nG.
    [5]
    5. Method (100) according to claims 1 to 4, wherein the definition (300) of the topographic code (360) of the determined structure (270) comprises a step of assigning (310) a coding element said principal (ECP) to each identified steered path (CH) according to principal characteristics (260) of the identified steered path (CH), the main features (260) being: - the nature of the vertices (T, C) and / or - the numbers nG, nD and the situation of the edges coming from the at least one intermediate crossing vertex (Cl).
    [6]
    The method (100) according to one of claims 1 to 5, wherein, when the representations of sets of graphs (110) are from skeletonized images, the definition (300) of the topographic code (360) of the determined structure (270) further comprises a step of assigning (315) an said appendix coding element (ECA) to each listed directed path (CH) according to related characteristics (265) of the identified directed path (CH) , the additional characteristics (265) being: the ratio (p) between the number of pixels of an edge of the listed directed path (CH) appearing on the graphical representation and the number of pixels joining in a straight line the two vertices of said edge, - for each intermediate intersection (Cl) of the identified oriented road (CH), the indication (y) of the angle formed by the direction of the edge of the intermediate intersection (C1) with respect to the direction of the ridge arriving at this intermedia crossing ire (Cl), - for each intermediate crossing point (Cl) of the listed directed path (CH), when the number nG; nD edges resulting from this intermediate crossing vertex (C1) is non-zero, the ratio existing between, on the one hand, the value (a) of the angles formed by the first edge encountered in each of the scanning directions with the arriving edge at the intermediate crossing vertex (C1) and, on the other hand, the value (a) of the angles formed by the last edge encountered in each of the scanning directions with the edge coming from the intermediate crossing vertex (C1).
    [7]
    7. Method (100) according to one of claims 1 to 6, wherein the comparison (400) of the defined topographic code (360) with the preset topographic code (390) further comprises a consideration in the elaboration (420) of the similarity measure, ancillary coding elements (ECA).
    [8]
    A computer program product comprising code instructions for executing a mapping method (100) according to one of the preceding claims when the program is executed by a processor.
    [9]
    9. System (500) for identifying and / or authenticating an individual for carrying out the mapping method (100) according to one of claims 1 to 9, comprising a production device ( 510), composed of a capture device (520) dedicated to the acquisition of digital images (525) in the visible domain and / or in the infrared domain and a data processing device (530 ) configured to process the digital images acquired by the capture device (520), being able to extract a skeletonized image (533) from the digital image (525), characterized in that it is furthermore capable of generating a set of planar graphs (110) from the skeletonized image (533) and executing on said set of graphs (110) determination (200) and definition (300) steps, and in that the system comprises in addition - an evaluation device (550), consisting of a device storage device (560) storing at least one preset topographic code (390) and a comparison device (570) able to develop (420) and qualify (420) a similarity measure between a defined binary code (360) provided by the production device (510) and a pre-established binary code (390) present in the storage device (560).
    [10]
    System (500) according to claim 9, characterized in that - the data processing device (530) comprises a computer and communication interfaces with the capture (520) and evaluation (550) devices, in what - the storage device (560) is realized using any type of memory - RAM, Flash, hard disk or SSD, that the binary codes are organized into databases, and in that - the capture (520), data processing (530), storage (560) and comparison (570) devices may comprise calculators and communication interfaces connecting them to each other.
    [11]
    11. System (500) according to claims 9 and 10, characterized in that: - the computers consist of processors or microprocessors, presented in the form of controllers, microcontrollers, DSP, FPGA or any other combination to implement the process steps (100) and, capable of adding to the topographic bit codes (360, 390) any relevant context information for the particular implementation, including a precise geographic position, GPS or equivalent and the date and time of capturing the digital image (525), and in that - the communication interfaces between devices are wired or wireless and use communication protocols such as RS-232, USB, FireWire, HDMI, Bluetooth, ZigBee, Ethernet, Wifi or mobile telephony, satellite or fixed.
    [12]
    12. System (500) according to one of claims 9 to 11 wherein the capture devices (520) and treatment (530) can be integrated within the same housing, such as a mobile phone.
    [13]
    13. Use of the method (100) of mapping representations of planar related graph sets (110) according to one of claims 1 to 7 in the field of biometric authentication, when these graphs (110) are derived from skeletonized images of venous networks and / or in the field of electronic mapping when these graphs (110) are derived from skeletonized aerial photographs.
    类似技术:
    公开号 | 公开日 | 专利标题
    Zampoglou et al.2017|Large-scale evaluation of splicing localization algorithms for web images
    Wu et al.2017|Deep matching and validation network: An end-to-end solution to constrained image splicing localization and detection
    WO2010023415A1|2010-03-04|Method of determining a pseudo-identity on the basis of characteristics of minutiae and associated device
    Skurikhin et al.2013|Automated tree crown detection and size estimation using multi-scale analysis of high-resolution satellite imagery
    Guo et al.2014|A multi-stage approach to curve extraction
    FR2979727A1|2013-03-08|IDENTIFICATION BY RECOGNITION OF IRIS
    CH713468A2|2018-08-15|A method of mapping sets of planar related graphs by topographic coding, system and associated uses.
    Juneja2017|A noise robust VDD composed PCA-LDA model for face recognition
    EP3620970A1|2020-03-11|Method for extracting characteristics of a fingerprint represented by an input image
    Nguyen et al.2020|Universal fingerprint minutiae extractor using convolutional neural networks
    Imdahl et al.2015|Towards generating realistic synthetic fingerprint images
    Debbarma et al.2014|Keypoints based copy-move forgery detection of digital images
    Wrobel et al.2016|Effective lip prints preprocessing and matching methods
    FR3088467A1|2020-05-15|METHOD FOR CLASSIFYING A REPRESENTATIVE INPUT IMAGE OF A BIOMETRIC TRAIT USING A CONVOLUTIONAL NEURON NETWORK
    Zhang et al.2016|Detection of copy-move forgery in flat region based on feature enhancement
    Mohanty et al.2019|A photo forensics-based prototype to combat revenge porn
    kumar Shukla et al.2016|A novel method for identification and performance improvement of Blurred and Noisy Images using modified facial deblur inference | algorithms
    Li et al.2017|Evaluation of super-resolution on bird detection performance based on deep convolutional networks
    Muzaffer et al.2015|Copy Move Forgery Detection Using Gabor Filter and ORB
    Mazumdar et al.2020|Siamese convolutional neural network-based approach towards universal image forensics
    Ellingsgaard et al.2017|Altered fingerprint detection
    Abdosalehi et al.2019|Splicing localization in tampered blurred images
    Wang et al.2010|Edge detection from high-resolution remotely sensed imagery based on Gabor filter in frequency domain
    Jindal et al.2019|Digital Image Forensics-Gateway to Authenticity: Crafted with Observations, Trends and Forecasts
    Arun Anoop et al.2019|Study of the importance of digital forensics and deep learning tools
    同族专利:
    公开号 | 公开日
    FR3062740A1|2018-08-10|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    法律状态:
    2021-03-15| AZW| Rejection (application)|
    优先权:
    申请号 | 申请日 | 专利标题
    FR1770117A|FR3062740A1|2017-02-03|2017-02-03|METHOD FOR MAPPING MAP-RELATED GRAPH RELATED SETS BY TOPOGRAPHIC CODING, SYSTEM AND USES THEREOF|
    [返回顶部]