Signature Content Recognition Method and System

Abstract

An embodiment of the present invention provides a signature content recognition method and system. The method includes: obtaining signature information to be authenticated; preprocessing the signature information to obtain feature information of acquisition points; extracting the feature information of feature points from the feature information of acquisition points based on the bending degree of the signature glyph curve; and matching the feature information of the feature points with a template library to identify whether the signature information to be authenticated passes the authentication. The embodiment of the present invention innovatively extracts the feature information of feature points from all the preprocessed feature information of acquisition points according to the bending degree of the signature glyph curve, and uses the method of obtaining feature points by curvature change or speed, reducing the influence of the threshold value on the accuracy of feature point extraction. The embodiment of the present invention does not require processes such as obtaining the importance base value between adjacent points, simplifies the feature extraction process, and can comprehensively consider the bending degree of the glyph curve in the signature.

Description

Technical Field

[0001] The present invention relates to the field of business support, and particularly to a signature content recognition method and system. Background Art

[0002] A signature refers to the handwritten vector data for a customer to confirm the business acceptance content in a business acceptance form generated during daily production work. Signature recognition refers to recognizing the customer signature data in the business acceptance form and identifying the signature content of the customer in the business acceptance form. As an important branch of biometric technology, handwritten signature authentication has the advantages of being difficult to imitate, easy to accept, and easy to collect information, and has always been a research hotspot in the field of biometric technology.

[0003] According to the relevant provisions of the "Electronic Signature Law", the customer signature information in the business acceptance form represents the customer's confirmation of the business acceptance content described in the business document and has certain legal effects. At present, in the business acceptance link, the content of the customer's signature information is not confirmed. Instead, the signature in the pdf is viewed manually, and the supervisor judges the signature content.

[0004] However, at present, in the business acceptance link, the content of the customer's signature information is not confirmed. Only whether there is a customer signature on the business acceptance form is judged through the business system. This verification is completed by calling the control API interface and cannot verify the signature content. Customers can write randomly, and only manual verification can be carried out through the later audit process. Summary of the Invention

[0005] The present invention provides a signature content recognition method and system to solve the technical defects existing in the prior art.

[0006] The present invention provides a signature content recognition method, including:

[0007] Obtaining the signature information to be authenticated;

[0008] Preprocessing the signature information to obtain the feature information of the acquisition points;

[0009] Extracting the feature information of the feature points from the feature information of the acquisition points based on the bending degree of the signature glyph curve;

[0010] Matching the feature information of the feature points with the template library to identify whether the signature information to be authenticated passes the authentication.

[0011] According to the signature content recognition method provided by the present invention, the obtaining of the signature information to be authenticated includes:

[0012] During the signature process, the position, time, pressure, the order of the signature strokes, as well as the speed and acceleration corresponding to the signature strokes of each signature point are obtained in real time.

[0013] According to the signature content recognition method provided by the present invention, the preprocessing of the signature information includes:

[0014] Preprocess the signature information by denoising, smoothing, and normalizing.

[0015] According to the signature content recognition method provided by the present invention, the extracting of the feature information of the feature points from the collected point feature information based on the bending degree of the signature glyph curve includes:

[0016] Use the following formula to obtain the bending degree of each signature point in the signature glyph curve, that is, the angle ∠Ai:

[0017]

[0018] Where the point Pi is the coordinate point (Xi, Yi) of the signature point after preprocessing, i ∈ [m + 1, N - m], m is a parameter, and N is the total number of collected signature points; ∠Ai is and The included angle, where 0° ≤ ∠Ai ≤ 180°;

[0019] Based on the bending degree of each signature point, obtain the angle curve of the signature glyph;

[0020] Perform the first derivative on the angle curve of the signature glyph, and use the signature points with the derivative less than zero and the corresponding angle between 0° and 145° as the feature information for extracting feature points.

[0021] According to the signature content recognition method provided by the present invention, before identifying whether the signature information to be authenticated passes the authentication based on matching the feature information of the feature points with the template library includes:

[0022] Judge whether the number of strokes in the feature information of the feature points is the same as that of the signatures existing in the template library and the number of strokes is the same. If not, it is identified that the signature information to be authenticated fails the authentication.

[0023] According to the signature content recognition method provided by the present invention, the matching of the feature information of the feature points with the signature features in the template library and identifying whether the signature information to be authenticated passes the authentication includes:

[0024] Calculate the similarity between the feature sequence in the feature information of the feature points and the feature sequence in the template library;

[0025] Judge whether there is a signature in the template library whose similarity with the feature points is greater than the threshold;

[0026] If so, it is identified that the signature information to be authenticated passes the authentication;

[0027] If not, it is recognized that the signature information to be authenticated fails the authentication.

[0028] According to the signature content recognition method provided by the present invention, the matching of the feature information of the feature points with the signature features in the template library and the recognition of whether the signature information to be authenticated passes the authentication include:

[0029] Based on the feature information of the feature points, determine the weight of each feature point;

[0030] Use the weights of each feature point and the similarity between each feature point to obtain a signature matching result.

[0031] The present invention also provides a signature content recognition system, including:

[0032] A signature information acquisition module, configured to acquire signature information to be authenticated;

[0033] A preprocessing module, configured to preprocess the signature information to obtain acquisition point feature information;

[0034] A feature extraction module, configured to extract the feature information of feature points from the acquisition point feature information based on the bending degree of the signature glyph curve;

[0035] An identification and authentication module, configured to match the feature information of the feature points with a template library to identify whether the signature information to be authenticated passes the authentication.

[0036] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of any one of the above signature content recognition methods are implemented.

[0037] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the above signature content recognition methods are implemented.

[0038] In the signature content recognition method of the embodiment of the present invention, innovatively, according to the bending degree of the signature glyph curve, the feature information of feature points is extracted from all the acquisition point feature information after preprocessing. By using the method of obtaining feature points by curvature change or speed, the feature information of the feature points is matched with a template library to identify whether the signature information to be authenticated passes the authentication, reducing the influence of the threshold value on the accuracy of feature point extraction. The embodiment of the present invention does not require processes such as obtaining the importance base value between adjacent points, simplifies the feature extraction process, and can comprehensively consider the bending degree of the glyph curve in the signature. Description of the Drawings

[0039] To more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0040] Figure 1 It is a schematic flow chart of a signature content recognition method provided by an embodiment of the present invention;

[0041] Figure 2 It is a schematic diagram of the signature font angle provided by an embodiment of the present invention;

[0042] Figure 3 It is a schematic diagram of a signature content recognition system provided by an embodiment of the present invention;

[0043] Figure 4 It is a schematic structural diagram of an electronic device provided by the present invention;

[0044] Figure 5 They are the feature points obtained by using the signature content recognition method provided by an embodiment of the present invention for the signature font. Detailed implementation manners

[0045] To make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention with reference to the accompanying drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present invention belong to the scope of protection of the present invention.

[0046] An embodiment of the present invention discloses a signature content recognition method. Refer to Figure 1 , including:

[0047] S1: Obtain the signature information to be authenticated;

[0048] The signature information includes the signature point position, stroke order, time information of each signature point, and information such as the pressure information, speed, and acceleration of the signature point. Among them, the signature point refers to multiple points corresponding to the writing strokes formed by writing on the touchpad during the handwritten signature process.

[0049] S2: Preprocess the signature information to obtain the feature information of the acquisition points;

[0050] S3: Extract the feature information of the feature points from the feature information of the acquisition points based on the bending degree of the signature font curve;

[0051] Specifically, according to the bending degree of the signature glyph curve, the feature information of the feature points is extracted from all the preprocessed collected point feature information. Compared with the prior art, the embodiments of the present invention use the method of obtaining feature points by curvature change (maximum curvature) or speed (minimum speed), which reduces the influence of the threshold value on the accuracy of feature point extraction. In addition, compared with the J.Bruatl feature point detection algorithm, the embodiments of the present invention do not require processes such as obtaining the importance base value between adjacent points, simplifying the feature extraction process, and the embodiments of the present invention can comprehensively consider the bending degree of the glyph curve in the signature.

[0052] S4: Match the feature information of the feature points with the template library to identify whether the signature information to be authenticated passes the authentication.

[0053] For signature matching authentication, the feature points are used to match with the signatures in the template library to verify whether the signature passes. If there is a signature in the template library that matches the feature information of the feature points, it is identified that the signature information to be authenticated passes the authentication. If there is no signature in the template library that matches the feature information of the feature points, it is identified that the signature information to be authenticated fails the authentication.

[0054] The embodiments of the present invention innovatively extract the feature information of the feature points from all the preprocessed collected point feature information according to the bending degree of the signature glyph curve, use the method of obtaining feature points by curvature change or speed, match the feature information of the feature points with the template library, and identify whether the signature information to be authenticated passes the authentication, reducing the influence of the threshold value on the accuracy of feature point extraction. The embodiments of the present invention do not require processes such as obtaining the importance base value between adjacent points, simplifying the feature extraction process, and can comprehensively consider the bending degree of the glyph curve in the signature.

[0055] According to the signature content recognition method provided by the present invention, the obtaining of the signature information to be authenticated includes:

[0056] During the signature process, the position, time, pressure, the order of the signature strokes, as well as the speed and acceleration corresponding to the signature strokes of each signature point are obtained in real time on the touchpad.

[0057] According to the signature content recognition method provided by the present invention, the preprocessing of the signature information includes:

[0058] The signature information is preprocessed by means of denoising, smoothing, and normalization.

[0059] According to the signature content recognition method provided by the present invention, the extracting of the feature information of the feature points from the collected point feature information based on the bending degree of the signature glyph curve includes:

[0060] The bending degree of each signature point in the signature glyph curve is obtained using the following formula, i.e., the angle ∠Ai:

[0061]

[0062] where, as Figure 2 shown, the point Pi is the coordinate point (Xi, Yi) of the signature point after preprocessing, i ∈ [m + 1, N - m], m is a parameter with a value of 3, and N is the total number of collected signature points; ∠Ai is the and included angle, where 0° ≤ ∠Ai ≤ 180°;

[0063] Based on the bending degree of each signature point, the angle curve of the signature glyph is obtained;

[0064] The obtained ∠Ai is described in a two-dimensional manner: using the ordinate of the signature point to represent ∠Ai and the abscissa to represent the time series of the signature point, the relevant angle curve of the glyph is obtained.

[0065] The angle curve of the signature glyph is differentiated once, and the signature points with a derivative less than zero and corresponding angles between 0° and 145° are used as the feature information for extracting feature points, where the feature points include the starting point and the ending point of the signature. Figure 5 Shows the feature points obtained by this method for some signature glyphs (the positions with dots are the obtained feature points).

[0066] According to the signature content recognition method provided by the present invention, before identifying whether the signature information to be authenticated passes the authentication based on matching the feature information of the feature points with the template library, it includes:

[0067] Judging whether the number of strokes in the feature information of the feature points is the same as that of the signature existing in the template library and the number of strokes is the same. If not, it is identified that the signature information to be authenticated fails the authentication.

[0068] That is to say, before S4 in the embodiment of the present invention, the obtained number of strokes of the signature can also be used to determine whether to perform step S4. If there is a signature template in the template library that is the same as the name of the signature and has the same number of strokes, then S4 is executed; otherwise, it is determined that the signature fails.

[0069] According to the signature content recognition method provided by the present invention, the matching of the feature information of the feature points with the signature features in the template library to identify whether the signature information to be authenticated passes the authentication includes:

[0070] Calculating the similarity between the feature sequence in the feature information of the feature points and the feature sequence in the template library;

[0071] Determine whether there is a signature in the signature template library whose similarity to the feature points is greater than the threshold;

[0072] If so, it is recognized that the signature information to be authenticated passes the authentication;

[0073] If not, it is recognized that the signature information to be authenticated fails the authentication.

[0074] Calculate the similarity between the feature sequence of the feature points of the signature and the feature sequence of the feature points of the signature template. If there is a signature in the template library whose similarity to the signature is greater than the preset threshold, the verification passes; otherwise, the verification fails. Or, if the average value of the similarity between the signature and the signatures in the template library is greater than the preset threshold, the verification passes; otherwise, the verification fails.

[0075] According to the signature content recognition method provided by the present invention, the matching of the feature information of the feature points with the signature features in the template library and the recognition of whether the signature information to be authenticated passes the authentication include:

[0076] Based on the feature information of the feature points, determine the weight of each feature point;

[0077] Use the weights of each feature point and the similarity between each feature point to obtain the signature matching result.

[0078] Embodiments of the present invention can configure corresponding weights for each feature point. Among them, the weight allocation is obtained based on each signature template with successful authentication, that is, the feature information of each signature template is fused, and the accuracy of feature matching is improved through the configuration of weights. The improved DTW algorithm is used to set adaptive weights for each feature point, and the similarity between the signature and the signature template is obtained by using the weights and the signature features.

[0079] That is to say, S4 includes:

[0080] S41: Determine the weight of the feature point;

[0081] 1) Set the initial weight of each feature point in the signature feature to w = 1 / m, (k = 1, 2,..., m), where m is the number of feature points.

[0082] 2) Calculate the similarity between the signature feature and the signature template;

[0083] 3) Calculate the similarity between the signature feature and the signature template when removing a certain corresponding feature point from the signature feature and the signature template; <>

[0084] 4) Use the difference between 2 and 3 to determine the weight of the removed feature point.

[0085] S42: Obtain the signature matching result by using the weights of the feature points and the similarity between each feature point in the template library.

[0086] Among them, the feature point matching algorithm that improves DTW adds the stability weighting coefficient wi of the feature point to be matched on the basis of the basic DTW feature point matching algorithm. That is, the matching algorithm between the sequence X of feature points of the signature to be authenticated and the sequence G of feature points of the template signature can be expressed by the dynamic programming method as follows:

[0087]

[0088] Among them, the value of DM(G, X) reflects the matching distance between the two sequences of signature feature points. If DM(G, X) is less than the threshold S, it is considered that the sequence of feature points is successfully matched, and thus it can be determined that the signature to be authenticated is a genuine signature; otherwise, it can be determined that the signature to be authenticated is a forged signature.

[0089] In the embodiment of the present invention, on the basis of comprehensively considering the coordinates, time, and feature point types (such as position, pressure, acceleration, etc.) of the feature points, a similarity function for measuring the difference between two feature points is defined. The method is to represent each feature point with a four-dimensional vector e(e(x), e(y), e(t), e(s)), where e(x) and e(y) represent the coordinates of the feature point on the x and y axes, e(t) represents the time of the feature point, and e(s) is the type of the feature point. G(g1, g2, g3,..., gm) and X(x1, x2, x3,..., x n ) respectively represent the sequence of feature points of the template signature and the signature to be authenticated. In this way, the similarity between two feature points can be expressed by the formula:

[0090] s ij = λ ij ·τ ij

[0091] Among them, τ ij and λ ij are respectively defined as follows:

[0092]

[0093] Among them, S e and S t are both constants, and satisfy the condition S e ≥ S t ≥ 0, e i (s), e j (s) represents the type of the feature point.

[0094]

[0095] Among them, x' i = t' i(x) / W, x′ i = t′ i (x) / W, y′ i = t′ i (x) / H, y′ j = t′ j (x) / H, t′ i = e i (t) / N,

[0096] t′ j = e j (T) / N, W is the maximum width of the handwritten signature in the X-axis direction, H is the maximum height of the signature, and N is the number of signature sample points. From this, an m×n matrix composed of sij (i ∈ (1, m), j ∈ (1, n)) can be obtained, where m and n are respectively equal to the number of feature points of the template sample and the sample to be authenticated. Thus, a similarity matrix D between any two feature points can be obtained, as shown in Table 2:

[0097] Table 2

[0098]

[0099] Each element in D represents the similarity between the sequence of feature points to be authenticated and the sequence of feature points of the signature sample in the template.

[0100] Among them, the process of finding matching points can be referred to as follows: Let the current decision point be D(i, j), indicating that the i-th feature point of the current signature to be authenticated matches the j-th feature point in the template signature; from the properties of the DTW matching algorithm, it can be known that the next matching point will be one of the three points D(i + 1, j), D(i + 1, j + 1), D(i, j + 1), and the selection is made according to the following rules:

[0101] If one of the adjacent elements D(i + 1, j), D(i + 1, j + 1), and D(i, j + 1) is less than the other two elements, the forward direction points to this element;

[0102] If D(i + 1, j) = D(i + 1, j + 1) and D(i + 1, j + 1) < D(i, j + 1), that is, the elements on the right and upper right are equal and less than the element on the upper side, then choose to go right;

[0103] If D(i, j + 1) = D(i + 1, j + 1) and D(i + 1, j + 1) < D(i + 1, j), that is, the elements on the upper side and upper right are equal and less than the element on the right side, then choose to go up;

[0104] If D(i + 1, j) = D(i, j + 1) and D(i, j + 1) < D(i + 1, j + 1), that is, the elements on the right and above are equal and less than the element in the upper right, then compare the second adjacent elements on the right and above. If they are still equal, then push the compared elements one step further to the right and up until two unequal elements are found. At this time, if the element on the right is smaller, then move to the right; otherwise, move upward;

[0105] If D(i + 1, j) = D(i + 1, j + 1) = D(i, j + 1), that is, three adjacent elements are equal. At this time, first determine whether the i-th feature point can be matched. If it can, then move to the upper left; otherwise, process it according to the method in (d).

[0106] Through this decision-making process, the recursive problem in the DTW algorithm can be eliminated, thereby saving a large amount of storage space and calculation time and improving the efficiency of the algorithm.

[0107] In addition, the embodiments of the present invention can also be used in the signature template generation process. The signature confirmed to pass the verification is stored in the template library, and signatures with the same name and the same number of strokes are stored in a group for the verification process of the number of strokes. The initial template library is pre-input, and during the application of the embodiments of the present invention, it can be continuously updated based on the signatures that pass the verification.

[0108] The embodiments of the present invention disclose a signature content recognition system, see Figure 3 , including:

[0109] A signature information acquisition module 10, configured to acquire signature information to be authenticated;

[0110] The signature information includes the signature point position, stroke order, time information of each signature point, and information such as the pressure information, speed, and acceleration of the signature point. Among them, the signature point refers to multiple points corresponding to the writing strokes formed by writing on the touchpad during the handwritten signature process.

[0111] A preprocessing module 20, configured to preprocess the signature information to obtain the feature information of the acquisition points;

[0112] A feature extraction module 30, configured to extract the feature information of the feature points from the feature information of the acquisition points based on the bending degree of the signature glyph curve;

[0113] Specifically, according to the bending degree of the signature glyph curve, the feature information of the feature points is extracted from all the preprocessed collected point feature information. Compared with the prior art, the method for obtaining feature points using curvature change (maximum curvature) or speed (minimum speed) in the embodiments of the present invention reduces the influence of the threshold value on the accuracy of feature point extraction. In addition, compared with the J.Bruatl feature point detection algorithm, the embodiments of the present invention do not require processes such as obtaining the importance base value between adjacent points, simplifying the feature extraction process, and the embodiments of the present invention can comprehensively consider the bending degree of the glyph curve in the signature.

[0114] An identification and authentication module 40 is configured to match the feature information of the feature points with a template library to identify whether the signature information to be authenticated passes the authentication.

[0115] For signature matching authentication, the feature points are used to match with the signatures in the template library to verify whether the signature passes. If there is a signature in the template library that matches the feature information of the feature points, it is identified that the signature information to be authenticated passes the authentication. If there is no signature in the template library that matches the feature information of the feature points, it is identified that the signature information to be authenticated does not pass the authentication.

[0116] Further, the signature information acquisition module 10 is configured to, during the signature process, real-time acquire the position, time, pressure, the order of the signature strokes, as well as the speed and acceleration corresponding to the signature strokes of each signature point on the touchpad.

[0117] The preprocessing module 20 is specifically configured to preprocess the signature information by means of denoising, smoothing, and normalization.

[0118] The extraction of the feature information of the feature points from the collected point feature information based on the bending degree of the signature glyph curve includes:

[0119] The bending degree of each signature point in the signature glyph curve, that is, the angle ∠Ai, is obtained by using the following formula:

[0120]

[0121] where the point Pi is the coordinate point (Xi, Yi) of the signature point after preprocessing, i ∈ [m + 1, N - m], m is a parameter with a value of 3, and N is the total number of collected signature points; ∠Ai is and the included angle, where 00 ≤ ∠Ai ≤ 1800;

[0122] Based on the bending degree of each signature point, an angle curve of the signature font is obtained;

[0123] Describe the obtained ∠Ai in a two-dimensional manner: use the ordinate of the signature point to represent ∠Ai, and use the abscissa to represent the time series of the signature points to obtain the relevant angle curve of the font.

[0124] Perform a first derivative on the angle curve of the signature font, and use the signature points where the derivative is less than zero and the corresponding angle is between 0° and 145° as the characteristic information for extracting characteristic points. The characteristic points include the starting point and the ending point of the signature.

[0125] The recognition and authentication module 40 is specifically used to calculate the similarity between the characteristic sequence in the characteristic information of the characteristic points and the characteristic sequence in the template library.

[0126] Determine whether there is a signature in the template library whose similarity with the characteristic points is greater than the threshold.

[0127] If so, it is recognized that the signature information to be authenticated passes the authentication.

[0128] If not, it is recognized that the signature information to be authenticated fails the authentication.

[0129] Calculate the similarity between the characteristic sequence of the characteristic points of the signature and the characteristic sequence of the characteristic points of the signature template. If there is a signature in the template library whose similarity with the signature is greater than the preset threshold, the verification passes; otherwise, the verification fails. Or, if the average value of the similarity between the signature and the signatures in the template library is greater than the preset threshold, the verification passes; otherwise, the verification fails.

[0130] The recognition and authentication module 40 is specifically used to determine the weight of each characteristic point based on the characteristic information of the characteristic points.

[0131] Use the weights of each characteristic point and the similarity between each characteristic point to obtain the signature matching result.

[0132] In the embodiment of the present invention, corresponding weights can be configured for each characteristic point. Among them, the weight allocation is obtained based on each signature template with successful authentication, that is, the characteristic information of each signature template is fused, and the accuracy of feature matching is improved through the configuration of weights. Use the improved DTW algorithm to set adaptive weights for each characteristic point, and use the weights and signature features to obtain the similarity between the signature and the signature template.

[0133] Figure 4The figure exemplifies a schematic diagram of the physical structure of an electronic device, which may include: a processor 310, a communications interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 complete communication with each other through the communication bus 340. The processor 310 can call the logical instructions in the memory 330 to execute a signature content recognition method, which includes:

[0134] S1: Obtain the signature information to be authenticated;

[0135] S2: Preprocess the signature information to obtain the feature information of the acquisition points;

[0136] S3: Extract the feature information of the feature points from the feature information of the acquisition points based on the bending degree of the signature glyph curve;

[0137] S4: Match the feature information of the feature points with the template library to identify whether the signature information to be authenticated passes the authentication.

[0138] In addition, when the logical instructions in the above-mentioned memory are implemented in the form of software functional units and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0139] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute a signature content recognition method, which includes:

[0140] S1: Obtain the signature information to be authenticated;

[0141] S2: Preprocess the signature information to obtain the feature information of the acquisition points;

[0142] S3: Extract the feature information of the feature points from the collected point feature information based on the degree of bending of the signature glyph curve;

[0143] S4: Match the feature information of the feature points with the template library to identify whether the signature information to be authenticated passes the authentication.

[0144] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is configured to execute a signature content recognition method, and the method includes:

[0145] S1: Obtain the signature information to be authenticated;

[0146] S2: Preprocess the signature information to obtain the collected point feature information;

[0147] S3: Extract the feature information of the feature points from the collected point feature information based on the degree of bending of the signature glyph curve;

[0148] S4: Match the feature information of the feature points with the template library to identify whether the signature information to be authenticated passes the authentication.

[0149] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0150] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A signature content recognition method, characterized in that, Including: Obtain the signature information to be authenticated; Preprocess the signature information to obtain the feature information of the collection points; Extract the feature information of the feature points from the feature information of the collection points based on the bending degree of the signature glyph curve; Match the feature information of the feature points with the signature features in the template library to identify whether the signature information to be authenticated passes the authentication; The matching of the feature information of the feature points with the signature features in the template library to identify whether the signature information to be authenticated passes the authentication includes: Determine the weight of each feature point based on the feature information of the feature points; Obtain the signature matching result by using the weights of each feature point and the similarity between each feature point; Among them, the determination of the weight of each feature point includes: S1: Set the initial weight of each feature point as w = 1 / m, where m is the number of feature points; S2: Calculate the similarity between the feature point and the signature feature; S3: Calculate the similarity between the feature point and the signature feature when removing a certain feature point and a corresponding signature feature in the template library; S4: Determine the weight of the removed feature point by using the difference between S2 and S3; The obtaining of the signature matching result by using the weights of each feature point and the similarity between each feature point includes: Use the improved DTW feature point matching algorithm to calculate the matching distance between the template signature feature sequence and the signature feature sequence to be authenticated; the improved DTW feature point matching algorithm adds the weight of the feature point to be matched on the basis of the basic DTW feature point matching algorithm; the template signature feature sequence is a sequence composed of the signature features in the template library; the signature feature sequence to be authenticated is a sequence composed of the feature points extracted from the signature information to be authenticated; the improved DTW feature point matching algorithm is represented by the following formula: ； where, i and j respectively represent the feature points extracted from the signature features in the template library and the signature information to be authenticated; G and X respectively represent the template signature feature sequence and the signature feature sequence to be authenticated; m represents the number of feature points in the template signature feature sequence; n represents the number of feature points in the signature feature sequence to be authenticated; x(t) represents the abscissa of the feature point, y(t) represents the ordinate of the feature point, and p(t) represents the coordinates of the feature point; w i represents the weight corresponding to the feature point; Judge whether the matching distance is less than a preset threshold; If so, identify that the signature information to be authenticated passes the authentication; If not, identify that the signature information to be authenticated does not pass the authentication.

2. The signature content recognition method according to claim 1, wherein The obtaining of the signature information to be authenticated includes: During the signature process, real-time obtain the position, time, pressure, the order of the signature strokes, as well as the speed and acceleration corresponding to the signature strokes of each signature point on the touchpad.

3. The signature content recognition method according to claim 1, wherein The preprocessing of the signature information includes: Preprocess the signature information by means of denoising, smoothing and normalization.

4. The signature content recognition method according to claim 3, wherein The extraction of the feature information of the feature points from the feature information of the collection points based on the bending degree of the signature glyph curve includes: Use the following formula to obtain the bending degree of each signature point in the signature glyph curve, that is, the angle ∠Ai: ； Among them, the point Pi is the coordinate point (Xi, Yi) of the signature point after preprocessing, i ∈ [m + 1, N - m], where m is a parameter and N is the total number of signature points collected; ∠Ai is and the included angle of, where 0° ≤ ∠Ai ≤ 180°; Obtain the angle curve of the signature glyph based on the bending degree of each signature point; Perform the first-order derivative on the angle curve of the signature glyph, and use the signature points with the derivative less than zero and the corresponding angle between 0° and 145° as the feature information of the extracted feature points.

5. The signature content recognition method according to claim 1, characterized in that Before identifying whether the signature information to be authenticated passes the authentication based on the matching of the feature information of the feature points with the template library: Determine whether the number of strokes in the feature information of the feature points is the same as that of the signature existing in the template library and the number of strokes is the same. If not, it is recognized that the signature information to be authenticated fails the authentication.

6. The signature content recognition method according to claim 1, characterized in that The matching of the feature information of the feature points with the signature features in the template library and the recognition of whether the signature information to be authenticated passes the authentication include: Calculate the similarity between the feature sequence in the feature information of the feature points and the feature sequence in the template library; Determine whether there is a signature in the template library whose similarity with the feature points is greater than the threshold; If so, it is recognized that the signature information to be authenticated passes the authentication; If not, it is recognized that the signature information to be authenticated fails the authentication.

7. A signature content recognition system, characterized in that, Include: A signature information acquisition module for acquiring signature information to be authenticated; A preprocessing module for preprocessing the signature information to obtain acquisition point feature information; A feature extraction module for extracting the feature information of feature points from the acquisition point feature information based on the bending degree of the signature glyph curve; An identification and authentication module for matching the feature information of the feature points with the signature features in the template library and identifying whether the signature information to be authenticated passes the authentication; The matching of the feature information of the feature points with the signature features in the template library and the recognition of whether the signature information to be authenticated passes the authentication include: Based on the feature information of the feature points, determine the weight of each feature point; Use the weights of each feature point and the similarity between each feature point to obtain a signature matching result; Among them, the determination of the weight of each feature point includes: S1: Set the initial weight of each feature point to w = 1 / m, where m is the number of feature points; S2: Calculate the similarity between the feature points and the signature features; S3: Calculate the similarity between the feature points and the signature features when removing a certain feature point and a corresponding signature feature in the template library; S4: Use the difference between S2 and S3 to determine the weight of the removed feature point; The use of the weights of each feature point and the similarity between each feature point to obtain a signature matching result includes: Use the improved DTW feature point matching algorithm to calculate the matching distance between the template signature feature sequence and the signature feature sequence to be authenticated; the improved DTW feature point matching algorithm adds the weight of the feature points to be matched on the basis of the basic DTW feature point matching algorithm; the template signature feature sequence is a sequence composed of the signature features in the template library; the signature feature sequence to be authenticated is a sequence composed of the feature points extracted from the signature information to be authenticated; the improved DTW feature point matching algorithm is represented by the following formula: ； Among them, i and j respectively represent the feature points extracted from the signature features in the template library and the signature information to be authenticated; G and X respectively represent the template signature feature sequence and the signature feature sequence to be authenticated; m represents the number of feature points in the template signature feature sequence; n represents the number of feature points in the signature feature sequence to be authenticated; x(t) represents the abscissa of the feature point, y(t) represents the ordinate of the feature point, and p(t) represents the coordinates of the feature point; w i represents the weight corresponding to the feature point; Determine whether the matching distance is less than the preset threshold; If so, it is recognized that the signature information to be authenticated passes the authentication; If not, it is recognized that the signature information to be authenticated fails the authentication.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, When the processor executes the program, it implements the steps of the signature content recognition method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the signature content recognition method according to any one of claims 1 to 6.

Images