Fingerprint recognition model training method, fingerprint recognition method, and related device

Abstract

Embodiments of the present application provide a fingerprint identification model training method, a fingerprint identification method and related equipment, which are used to solve the problem of inaccurate identification in fingerprint identification, so as to improve the fingerprint matching speed as much as possible. The fingerprint identification model training method in the embodiments of the present application comprises: obtaining a fingerprint library, wherein the fingerprint library comprises a to-be-matched fingerprint and a to-be-matched fingerprint rule corresponding to the to-be-matched fingerprint, the to-be-matched fingerprint rule comprises a to-be-matched rule string and a rule relationship between different rule strings; generating a fingerprint syntax tree according to the to-be-matched fingerprint and the rule relationship, and constructing a string matching model according to the to-be-matched rule string and the to-be-matched fingerprint; generating the fingerprint identification model according to the fingerprint syntax tree and the string matching model, so as to identify fingerprints according to the fingerprint identification model.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of computer fingerprint identification, and in particular to a fingerprint identification method and related equipment. BACKGROUND

[0002] When a user accesses a website or a device, some special scripts are used to collect user information, and the information is a fingerprint. The fingerprint can be further divided into website fingerprint and device fingerprint. No matter what kind of fingerprint is, the corresponding website or device can be found through the fingerprint.

[0003] In the existing technical means, the following shortcomings still exist in the identification through the fingerprint. For example, there are many formats of the fingerprint, and the differences between different fingerprint formats are large, causing each set of fingerprint to maintain a corresponding set of fingerprint identification engine. The separate fingerprint identification engine is not accurate for fingerprint identification of multiple formats, which increases the operating cost. SUMMARY

[0004] Embodiments of the present application provide a fingerprint identification model training method, a fingerprint identification method and related equipment, which are used to solve the problem of inaccurate identification in fingerprint identification, so as to improve the fingerprint matching speed as much as possible.

[0005] The first aspect of the embodiments of the present application provides a fingerprint identification model training method, comprising:

[0006] Obtaining a fingerprint library, wherein the fingerprint library comprises a to-be-matched fingerprint and a to-be-matched fingerprint rule corresponding to the to-be-matched fingerprint, the to-be-matched fingerprint rule comprising a to-be-matched rule string and a rule relationship between different rule strings;

[0007] Generating a fingerprint syntax tree according to the to-be-matched fingerprint and the rule relationship, and constructing a string matching model according to the to-be-matched rule string and the to-be-matched fingerprint;

[0008] Generating the fingerprint identification model according to the fingerprint syntax tree and the string matching model, so as to identify the fingerprint according to the fingerprint identification model.

[0009] Optionally, the generating the fingerprint syntax tree according to the to-be-matched fingerprint and the rule relationship comprises:

[0010] Repositioning the rule string with the rule relationship to form the rule string with a target rule relationship;

[0011] Positioning the rule string and the to-be-matched fingerprint according to the target rule relationship to form the fingerprint syntax tree.

[0012] Optionally, the constructing a string matching model according to the rule string to be matched and the fingerprint to be matched comprises:

[0013] Based on different fingerprints to be matched, the rule string to be matched corresponding to the fingerprint to be matched is taken as a node to construct an ordered tree;

[0014] Matching the fingerprint to be matched and the rule string to be matched in the ordered tree to obtain matching success information;

[0015] Constructing the string matching model according to the matching success information.

[0016] The second aspect of the embodiment of the present application provides a fingerprint identification method, comprising:

[0017] Obtaining a sample to be identified, wherein the sample to be identified comprises a fingerprint to be identified and a rule of the fingerprint to be identified corresponding to the fingerprint to be identified;

[0018] Inputting the fingerprint to be identified and the rule of the fingerprint to be identified into a pre-trained fingerprint identification model to determine whether the fingerprint to be identified is a target fingerprint satisfying a target rule of the fingerprint according to the fingerprint identification model; wherein the fingerprint identification model is generated by the training method of the first aspect.

[0019] Optionally, the fingerprint identification model comprises a string matching model and a fingerprint syntax tree, and the inputting the fingerprint to be identified and the rule of the fingerprint to be identified into the pre-trained fingerprint identification model to determine whether the fingerprint to be identified is a target fingerprint satisfying a target rule of the fingerprint according to the fingerprint identification model comprises:

[0020] Inputting all the rules of the fingerprint to be identified into the string matching model and matching all the rules of the fingerprint to be identified according to the string matching model to determine the rule of the fingerprint to be identified satisfying the target rule of the fingerprint;

[0021] If the rule of the fingerprint to be identified satisfies the target rule of the fingerprint, inputting the rule of the fingerprint to be identified satisfying the target rule of the fingerprint and the fingerprint to be identified into the fingerprint syntax tree and matching the fingerprint to be identified according to the fingerprint syntax tree to determine whether the fingerprint to be identified matches the target fingerprint in the fingerprint syntax tree.

[0022] Optionally, the method further comprises:

[0023] If one of the rules of the fingerprint to be identified does not satisfy the target rule of the fingerprint, determining that the fingerprint to be identified is an abnormal fingerprint, and matching the next rule of the fingerprint to be identified until all the rules of the fingerprint to be identified are matched.

[0024] Optionally, after judging whether the to-be-identified fingerprint is a target fingerprint satisfying a target fingerprint rule according to the fingerprint identification model, the method further comprises:

[0025] If the to-be-identified fingerprint does not match the target fingerprint, it is determined that the to-be-identified fingerprint is an abnormal fingerprint, and the next to-be-identified fingerprint is matched until all the to-be-identified fingerprints are matched.

[0026] Optionally, after judging whether the to-be-identified fingerprint is a target fingerprint satisfying a target fingerprint rule according to the fingerprint identification model, the method further comprises:

[0027] If the to-be-identified fingerprint is the target fingerprint satisfying the target fingerprint rule, the target fingerprint is matched according to a regular matching model to determine a fingerprint version type of the target fingerprint.

[0028] The third aspect of the embodiments of the present application provides a training system of a fingerprint identification model, comprising:

[0029] An acquisition unit is configured to acquire a fingerprint library, wherein the fingerprint library comprises a to-be-matched fingerprint and a to-be-matched fingerprint rule corresponding to the to-be-matched fingerprint, and the to-be-matched fingerprint rule comprises a to-be-matched rule string and a target rule relationship between different rule strings.

[0030] A generation unit is configured to generate a fingerprint syntax tree according to the to-be-matched fingerprint and the target rule relationship, and construct a string matching model according to the to-be-matched rule string and the to-be-matched fingerprint.

[0031] The generation unit is further configured to generate the fingerprint identification model according to the fingerprint syntax tree and the string matching model, so as to identify a fingerprint according to the fingerprint identification model.

[0032] Optionally, the training system of the fingerprint identification model further comprises a ranking unit.

[0033] The ranking unit is configured to rank the rule strings ranked according to the rule relationship, so as to form the rule strings having the target rule relationship.

[0034] The ranking unit is further configured to rank the rule strings and the to-be-matched fingerprint according to the target rule relationship, so as to form the fingerprint syntax tree.

[0035] Optionally, the training system of the fingerprint identification model further comprises a construction unit and a matching unit.

[0036] The constructing unit is configured to construct an ordered tree by taking the rule string to be matched corresponding to the fingerprint to be matched as a node based on different fingerprints to be matched.

[0037] The matching unit is configured to match the fingerprint to be matched and the rule string to be matched in the ordered tree to obtain matching success information.

[0038] The constructing unit is further configured to construct the string matching model according to the matching success information.

[0039] The third aspect of the embodiment of the present application provides a training method for executing the fingerprint identification model of the first aspect.

[0040] The fourth aspect of the embodiment of the present application provides a fingerprint identification system, comprising:

[0041] The obtaining unit is configured to obtain a sample to be identified, wherein the sample to be identified comprises a fingerprint to be identified and a rule of the fingerprint to be identified corresponding to the fingerprint to be identified.

[0042] The input unit is configured to input the fingerprint to be identified and the rule of the fingerprint to be identified into a pre-trained fingerprint identification model to determine whether the fingerprint to be identified is a target fingerprint satisfying a target rule of the fingerprint to be identified according to the fingerprint identification model, wherein the fingerprint identification model is generated by the training method of the first aspect.

[0043] Optionally, the fingerprint identification model comprises a string matching model and a fingerprint syntax tree, and the fingerprint identification comprises:

[0044] The input unit is specifically configured to input all the rules of the fingerprint to be identified into the string matching model, and match all the rules of the fingerprint to be identified according to the string matching model to determine a rule of the fingerprint to be identified satisfying the target rule of the fingerprint.

[0045] The input unit is further configured to input the rule of the fingerprint to be identified satisfying the target rule of the fingerprint and the fingerprint to be identified into the fingerprint syntax tree when the rule of the fingerprint to be identified satisfies the target rule of the fingerprint, and match the fingerprint to be identified according to the fingerprint syntax tree to determine whether the fingerprint to be identified matches the target fingerprint in the fingerprint syntax tree.

[0046] Optionally, the fingerprint identification system further comprises a determining unit.

[0047] The determination unit is configured to determine the to-be-identified fingerprint as an abnormal fingerprint when one of the to-be-identified fingerprint rules does not satisfy the target fingerprint rule, and perform matching on a next to-be-identified fingerprint rule until all the to-be-identified fingerprint rules are matched.

[0048] Optionally, the fingerprint identification system comprises:

[0049] The determination unit is further configured to determine the to-be-identified fingerprint as an abnormal fingerprint when the to-be-identified fingerprint does not match the target fingerprint, and perform matching on a next to-be-identified fingerprint until all the to-be-identified fingerprints are matched.

[0050] Optionally, the fingerprint identification system further comprises a matching unit.

[0051] The matching unit is configured to perform matching on the target fingerprint according to a regular matching model when the to-be-identified fingerprint is the target fingerprint satisfying the target fingerprint rule, to determine a fingerprint version type of the target fingerprint.

[0052] The fourth aspect of the embodiments of the present application provides a fingerprint identification method for performing the fingerprint identification method of the second aspect.

[0053] The fifth aspect of the embodiments of the present application provides a training device of a fingerprint identification model, comprising:

[0054] A central processing unit, a memory, an input and output interface, a wired or wireless network interface, and a power supply;

[0055] The memory is a transitory storage memory or a persistent storage memory;

[0056] The central processing unit is configured to communicate with the memory and perform instruction operations in the memory to execute the training method of the fingerprint identification model of the first aspect.

[0057] The sixth aspect of the embodiments of the present application provides a fingerprint identification device, comprising:

[0058] A central processing unit, a memory, an input and output interface, a wired or wireless network interface, and a power supply;

[0059] The memory is a transitory storage memory or a persistent storage memory;

[0060] The central processing unit is configured to communicate with the memory and perform instruction operations in the memory to execute the fingerprint identification method of the second aspect.

[0061] The seventh aspect of the embodiments of the present application provides a computer readable storage medium, characterized in that the computer readable storage medium comprises instructions, when the instructions are executed on a computer, the computer executes the method of the first aspect or the second aspect.

[0062] From the above technical solutions, the embodiments of the present application have the following advantages: through the fingerprint identification method disclosed by the embodiments of the present application, the to-be-identified sample is first obtained, wherein the to-be-identified sample includes a to-be-identified fingerprint and a to-be-identified fingerprint rule corresponding to the to-be-identified fingerprint; then the to-be-identified fingerprint and the to-be-identified fingerprint rule are input into a pre-trained fingerprint identification model to determine whether the to-be-identified fingerprint is a target fingerprint meeting a target fingerprint rule according to the fingerprint identification model. Thus, the to-be-identified fingerprint in the to-be-identified sample can be determined by the fingerprint identification model to obtain the target fingerprint meeting multiple conditions. Moreover, the same identification algorithm can be used to determine different to-be-identified fingerprints, thereby reducing the operating cost as much as possible.

[0063] Meanwhile, through the training method of the fingerprint identification model disclosed by the embodiments of the present application, a fingerprint library is first obtained, wherein the fingerprint library includes a to-be-matched fingerprint and a to-be-matched fingerprint rule corresponding to the to-be-matched fingerprint, the to-be-matched fingerprint rule includes a to-be-matched rule string and a rule relationship between different to-be-matched fingerprints; then a fingerprint syntax tree is generated according to the to-be-matched fingerprint and the rule relationship, and a string matching model is constructed according to the to-be-matched rule string and the to-be-matched fingerprint; finally, a fingerprint identification model is generated according to the fingerprint syntax tree and the string matching model to identify the fingerprint according to the fingerprint identification model. Thus, through the trained fingerprint identification model of the present application, the fingerprint to be identified can be effectively identified to solve the problem of slow and inaccurate fingerprint identification as much as possible. BRIEF DESCRIPTION OF DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art according to these drawings.

[0065] Figure 1 The flowchart of the training method of the fingerprint identification model disclosed by the embodiments of the present application is shown in the figure;

[0066] Figure 2 The flowchart of another training method of the fingerprint identification model disclosed by the embodiments of the present application is shown in the figure;

[0067] Figure 3 The flowchart of the fingerprint identification method disclosed by the embodiments of the present application is shown in the figure;

[0068] Figure 4 A flowchart of another fingerprint identification method disclosed in embodiments of the present application;

[0069] Figure 5 A structure diagram of a syntax tree disclosed in embodiments of the present application;

[0070] Figure 6 A structure diagram of a string matching model disclosed in embodiments of the present application;

[0071] Figure 7 A structure diagram of a training system of a fingerprint identification model disclosed in embodiments of the present application;

[0072] Figure 8 A structure diagram of a fingerprint identification system disclosed in embodiments of the present application;

[0073] Figure 9 A structure diagram of a training device of a fingerprint identification model disclosed in embodiments of the present application;

[0074] Figure 10 A structure diagram of a fingerprint identification device disclosed in embodiments of the present application. DETAILED DESCRIPTION

[0075] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0076] Current fingerprint identification has multiple shortcomings.

[0077] For example, first, the fingerprint format is not unified. There are currently multiple fingerprint formats, and the differences are very large. Each set of fingerprints must maintain a corresponding set of fingerprint matching engines. Second, the matching speed is slow: the current sequential matching method is adopted, and a large number of regular matching rules are included, and the time is completely determined by the number of fingerprints and the speed of regular matching. When the number of fingerprints reaches a certain level, the matching speed is very slow. Third, the fingerprint rule support is not perfect: the current fingerprint rule support is relatively single and cannot meet the complex rules of multiple condition dependencies.

[0078] To solve the above-mentioned problems of current fingerprint identification, such as non-uniform fingerprint format, slow matching speed, and imperfect fingerprint rule support, embodiments of the present application disclose a training method of a fingerprint identification model and related fingerprint identification methods. For details, please refer to the detailed description of the technical solutions of the present application.

[0079] Please refer toFigure 1 , Figure 1 A flowchart of a training method of a fingerprint identification model disclosed in an embodiment of the present application is shown. It includes steps 101-103.

[0080] 101, Obtain a fingerprint library.

[0081] Before training the fingerprint identification model, the fingerprint used for training needs to be obtained first.

[0082] In one specific embodiment, the fingerprint is trained by using a fingerprint identification engine. Specifically, the fingerprint identification engine can first obtain a fingerprint library. The fingerprint library is a collection of fingerprints, that is, a collection of records of commonly used fingerprints. In this embodiment, the fingerprint library includes a matching fingerprint and a matching fingerprint rule corresponding to the matching fingerprint. At the same time, the matching fingerprint rule includes a matching rule string and a rule relationship between different rule strings.

[0083] Based on the above embodiment, the fingerprint can specifically include a website fingerprint and / or a device fingerprint. The website fingerprint refers to the application name, version, front-end framework, back-end framework, server language, server operating system, website container, content management system, and / or database of the website. The device fingerprint includes the application name, version, open port, operating system, service name, and / or geographic location. The content contained in the fingerprint is not specifically limited, and will not be described in detail hereinafter.

[0084] Correspondingly, the fingerprint rule can be a string, and the rule relationship can be the relationship between the fingerprints. For example, in one specific embodiment, a fingerprint is listed as follows:

[0085] Finger: "header='test1' AND body='test2'"

[0086] Among them, header='test1' is a rule string, and body='test2' is also a rule string, and AND is a rule relationship. It is not difficult to understand that the above is only a specific example of one kind of fingerprint. For convenience of description, it will not be described in detail hereinafter. It should also be understood that based on the above embodiment, the fingerprint rule can support AND, OR, NOT (! =), priority (), and other rules. The rule relationship between the rule strings is not specifically limited here.

[0087] It should be noted that the fingerprints in the fingerprint library can be understood as fingerprints with multiple fingerprint formats and supporting multiple fingerprint rules.

[0088] 102. generating a fingerprint syntax tree according to the to-be-matched fingerprints and the rule relationship, and constructing a string matching model according to the to-be-matched rule string and the to-be-matched fingerprints.

[0089] In the process of loading the fingerprint library, the fingerprint recognition engine can generate a fingerprint syntax tree according to the to-be-matched fingerprints and the rule relationship in the fingerprint library. Meanwhile, a string matching model can also be constructed according to the to-be-matched rule string and the to-be-matched fingerprints.

[0090] In one specific embodiment, for the generation of the fingerprint syntax tree, the corresponding syntax tree is generated by using the loaded fingerprints until all the to-be-matched fingerprints are loaded. The plurality of fingerprint syntax trees can constitute a fingerprint syntax tree group.

[0091] In another specific embodiment, for the generation of the string matching model, the to-be-matched rule string is matched in advance, so as to construct a string matching model by using the matching results of different to-be-matched fingerprints, and to reduce the matching times of the pattern string and the main string by using the information after the failure, thereby achieving the purpose of fast matching.

[0092] 103. generating a fingerprint recognition model according to the fingerprint syntax tree and the string matching model, so as to recognize the fingerprint according to the fingerprint recognition model.

[0093] When the fingerprint syntax tree and the string matching model are generated, the fingerprint recognition engine generates a corresponding fingerprint recognition model. Thus, the fingerprint recognition model can be trained based on the to-be-matched fingerprints and the to-be-matched fingerprint rules in the loaded fingerprint library. Therefore, when the fingerprint recognition model receives a fingerprint to be recognized in the subsequent process, the to-be-recognized fingerprint can be recognized. For details of the fingerprint recognition process, please refer to the detailed description of the embodiments in Figure 3 or Figure 4 .

[0094] According to the training method of the fingerprint recognition model disclosed in the embodiment, the fingerprint library is first acquired, wherein the fingerprint library includes to-be-matched fingerprints and to-be-matched fingerprint rules corresponding to the to-be-matched fingerprints, the to-be-matched fingerprint rules include to-be-matched rule strings and rule relationships between different to-be-matched fingerprints; then a fingerprint syntax tree is generated according to the to-be-matched fingerprints and the rule relationship, and a string matching model is constructed according to the to-be-matched rule string and the to-be-matched fingerprints; finally, a fingerprint recognition model is generated according to the fingerprint syntax tree and the string matching model, so as to recognize the fingerprint according to the fingerprint recognition model. Thus, the fingerprint recognition model trained by the present application can effectively recognize the fingerprint to be recognized, so as to solve the problems of slow and inaccurate fingerprint recognition as much as possible.

[0095] For the convenience of the above Figure 1The training method of the fingerprint recognition model is described in detail in the foregoing method for training a fingerprint recognition model, and please refer to Figure 2 , Figure 2 The flowchart of another training method of a fingerprint recognition model disclosed in the embodiment of the present application is shown in FIG. 2. The method comprises steps 201-206.

[0096] 201. Obtain a fingerprint library.

[0097] Step 201 in the embodiment is similar to step 101 in the foregoing method for training a fingerprint recognition model, and will not be described in detail here. However, it should be noted that after all the fingerprints in the fingerprint library are obtained, the program corresponding to the fingerprint library loading will traverse the fingerprint library and load the fingerprints. Meanwhile, step 202 or step 204 will be performed during the fingerprint loading process. Details will be described later. Figure 1 202. Reorder the rule strings with regular relationships to form rule strings with target regular relationships.

[0098] Based on step 201, steps 202-203 are mainly for generating a fingerprint syntax tree. Specifically, the fingerprint recognition engine generates a corresponding syntax tree using the loaded fingerprints until the loading is completed. A plurality of fingerprint syntax trees constitute a fingerprint syntax tree group.

[0099] In one specific embodiment, the process of generating a syntax tree is to reorder the rule strings with regular relationships to form rule strings with target regular relationships by using a scheduling field algorithm. Based on this embodiment, the specific scheduling field algorithm is to convert the infix expressions in the fingerprint rules into reverse Polish expressions. Among them, the infix notation (or infix notation) is a general arithmetic or logical formula representation method, and the operator is in the middle of the operands in infix form (for example: 3+4). And the reverse Polish notation (or reverse Polish notation) is a method of placing all operators behind the operands, so it is also called postfix notation. The reverse Polish notation does not need parentheses to identify the priority of the operator.

[0100] 203. Arrange the rule strings and the to-be-matched fingerprints according to the target regular relationships to form a fingerprint syntax tree.

[0101] Based on step 202, when the target regular relationships of the rule strings are obtained, the rule strings and the to-be-matched fingerprints can be arranged according to the target regular relationships to form a fingerprint syntax tree.

[0102]

[0103] ​In one specific embodiment, the rule string and the to-be-matched fingerprint are sorted by the determined target rule relationship, so as to form a set of fingerprint syntax trees. Corresponding to a plurality of rule strings or to-be-matched fingerprints, a plurality of sets of fingerprint syntax trees are generated, so as to form a set of fingerprint syntax trees. For the convenience of describing the target rule relationship and the syntax tree, please refer to Figure 5 , Figure 5 A structural diagram of a syntax tree disclosed in an embodiment of the present application. For Figure 5 , it can be understood as a syntax tree generated by a to-be-matched fingerprint of (1+2)*3 and a rule relationship. The numbers 1, 2, and 3 can be understood as to-be-matched fingerprints. The “()”, “+”, and “*” can be understood as rule strings, and the corresponding ranking relationship is the rule relationship. For the convenience of description, this will not be described in detail hereinafter.

[0104] 204. Based on different to-be-matched fingerprints, the to-be-matched rule string corresponding to the to-be-matched fingerprint is taken as a node to construct an ordered tree.

[0105] Based on step 201, steps 204 to 205 are mainly for constructing a string matching model. Specifically, the to-be-matched rule string is added to the AC automaton (Aho-Corasick automaton) while the syntax tree is generated, until the fingerprint library is loaded. Based on the above embodiment, different AC automata can be established according to different to-be-matched fingerprints to form an AC automaton group. Taking a web fingerprint as an example, a plurality of AC automata such as head, body, and cert are established to form an AC automaton group. It is not difficult to understand that the above does not limit the construction method of the AC automaton group, and correspondingly, the AC automaton group can be understood as the string matching model described in the embodiment.

[0106] In combination with the above embodiment, it needs to be supplemented that the AC automaton belongs to an Aho-Corasick algorithm in computer science. The algorithm is used to match the sub-strings in the finite group “dictionary” in the input string. The difference between it and the ordinary string matching is that it matches all the dictionary strings at the same time. The algorithm has a linear time complexity in the average case, which is approximately the length of the string plus the number of all matches. However, since all matches need to be found, if each sub-string matches each other (for example, the dictionary is a, aa, aaa, aaaa, and the input string is aaaa), the time complexity of the algorithm will be approximately a quadratic function of the number of matches. For the convenience of description, the AC automaton will not be described hereinafter.

[0107] Since the AC automaton mainly runs the KMP (Knuth-Morris-Pratt) algorithm on the Trie tree for fast matching, it is easy to understand that in one specific embodiment, the fingerprint recognition engine constructs an ordered tree by using the rule strings corresponding to different fingerprints to be matched as nodes.

[0108] Based on the above embodiments, an ordered tree trie, also known as a prefix tree or dictionary tree, is used to store associative arrays, where the keys are typically strings. Unlike binary search trees, keys are not directly stored in nodes; instead, they are determined by the node's position in the tree. All descendants of a node share the same prefix, which is the string corresponding to that node, while the root node corresponds to an empty string. Generally, not all nodes have corresponding values; only the keys of leaf nodes and some internal nodes have relevant values.

[0109] The KMP algorithm (Knut-Morris-Platform operation) is an improved string matching algorithm. The core of the KMP algorithm is to utilize information from failed matches to minimize the number of matches between the pattern string and the main string, thus achieving fast matching. This is implemented through a `next()` function, which itself contains local matching information of the pattern string. Furthermore, the time complexity of the KMP algorithm is O(m+n).

[0110] based on Figure 1 The descriptions of fingerprints, rule relationships, and rule strings in the text are as follows: For different AC automata, header='test1' is the rule string, and test1 is built on the header automaton. Similarly, body='test2' is also a rule string, and test2 is built on the body automaton. For ease of description, these will not be elaborated further.

[0111] 205. Match the fingerprints to be matched and the rule strings to be matched in the ordered tree, obtain the matching success information, and construct a string matching model based on the matching success information.

[0112] Based on step 204, the fingerprints to be matched and the rule strings to be matched in the ordered tree are pre-matched to obtain matching success information and related matching failure information. The matching success information can be understood as the fingerprints to be matched and the rule strings forming a complete string, while the matching failure information can be understood as the fingerprints to be matched and the rule strings not forming a complete string.

[0113] For a detailed description of the matching process, please refer to [link / reference]. Figure 6 , Figure 6A structure diagram of a string matching model disclosed in an embodiment of the present application. For Figure 6 For {he}, it can be 012 or 34. However, if the combination with the least time consumption is to be obtained, 012 can be preferred. For the convenience of description, this will not be described in detail hereinafter.

[0114] 206. Generating a fingerprint identification model according to the fingerprint syntax tree and the string matching model, so as to identify the fingerprint according to the fingerprint identification model.

[0115] The step 206 in the embodiment is similar to the step 103 in the foregoing Figure 1 embodiment and will not be described in detail here.

[0116] The training method of the fingerprint identification model disclosed in the embodiment can replace the original sequential fingerprint matching with the AC automaton, so as to reduce the time complexity from O(N*M) to O(M+N), greatly improving the rule matching speed. Meanwhile, the fingerprint identification model disclosed in the embodiment can support various and complex fingerprint rules and also support fingerprints with various formats, thereby reducing the operating cost as much as possible.

[0117] Please refer to Figure 3 , Figure 3 A flowchart of a fingerprint identification method disclosed in an embodiment of the present application. It comprises steps 301-302.

[0118] 301. Obtaining a sample to be identified.

[0119] To identify the fingerprint, the sample to be identified is first obtained. The sample to be identified comprises a fingerprint to be identified and a fingerprint rule corresponding to the fingerprint to be identified.

[0120] In one specific embodiment, the sample to be identified can be an HTML page or other text, and the content contained in the sample to be identified is not limited here.

[0121] 302. Inputting the fingerprint to be identified and the fingerprint rule to be identified into a pre-trained fingerprint identification model, so as to determine whether the fingerprint to be identified is a target fingerprint satisfying a target fingerprint rule according to the fingerprint identification model.

[0122] When the to-be-identified fingerprint and the to-be-identified fingerprint rule are acquired, the to-be-identified fingerprint and the to-be-identified fingerprint rule acquired in the above steps can be input to the fingerprint identification model, so that the fingerprint identification model can determine whether the to-be-identified fingerprint is a target fingerprint meeting the target fingerprint rule.

[0123] In one specific embodiment, the fingerprint identification model described in the embodiment can be trained by the training method described in Figure 1 or Figure 2 The specific training steps are not described herein.

[0124] Based on the above embodiment, in another specific embodiment, the fingerprint identification model contains fingerprints of multiple formats and also stores multiple fingerprint rules. Therefore, in the embodiment, the target fingerprint rule and the target fingerprint (wherein the fingerprint library needs to contain the target fingerprint rule and the target fingerprint) can be set in advance, and it can also be determined whether the to-be-identified fingerprint or the to-be-identified fingerprint rule matches the fingerprint or the fingerprint rule in the fingerprint library.

[0125] By the fingerprint identification method disclosed in the embodiment, a to-be-identified sample is first acquired, wherein the to-be-identified sample includes a to-be-identified fingerprint and a to-be-identified fingerprint rule corresponding to the to-be-identified fingerprint; then the to-be-identified fingerprint and the to-be-identified fingerprint rule are input to a pre-trained fingerprint identification model to determine whether the to-be-identified fingerprint is a target fingerprint meeting a target fingerprint rule according to the fingerprint identification model. Thus, the to-be-identified fingerprint in the to-be-identified sample can be determined by the fingerprint identification model to obtain a target fingerprint meeting multiple conditions. Moreover, different to-be-identified fingerprints can be determined by using the same identification algorithm, thereby reducing the operating cost as much as possible.

[0126] For the convenience of describing the fingerprint identification method disclosed in the above Figure 3 in detail, please refer to Figure 4 , Figure 4 FIG. 2 is a flowchart of another fingerprint identification method disclosed in the embodiment. The method includes steps 401-406.

[0127] 401. Acquire a to-be-identified sample.

[0128] Step 401 in the embodiment is similar to step 301 in the above Figure 3 , and thus is not described herein.

[0129] 402. Input all to-be-identified fingerprint rules to a string matching model, and match all to-be-identified fingerprint rules according to the string matching model to determine a to-be-identified fingerprint rule meeting a target fingerprint rule.

[0130] It should be noted that the string matching model mentioned in the embodiment is aFigure 1 to Figure 2 In the embodiment, the fingerprint rule is determined preferentially. Specifically, all the to-be-identified fingerprint rules are input into the string matching model, so that the string matching model matches all the input to-be-identified fingerprint rules, and determines the to-be-identified fingerprint rule satisfying the target fingerprint rule.

[0131] In one specific embodiment, the AC automaton group is used to match the to-be-identified fingerprint rule by multiple AC automata, so as to determine whether the to-be-identified fingerprint rule hits the fingerprint rule in the fingerprint library. For example, the AC automaton is used to determine whether the to-be-identified fingerprint rule hits the fingerprint rule in the fingerprint library. For details, refer to step 204 in the method. Figure 2 For example, it can be determined whether the to-be-identified fingerprint rule hits header='test1'. It can be understood that the above is only one way of determining the to-be-identified fingerprint rule, and other matching methods can be set for other to-be-identified fingerprint rules to determine.

[0132] 403. If the to-be-identified fingerprint rule satisfies the target fingerprint rule, the to-be-identified fingerprint rule satisfying the target fingerprint rule and the to-be-identified fingerprint are input into the fingerprint syntax tree, and the to-be-identified fingerprint is matched according to the fingerprint syntax tree to determine whether the to-be-identified fingerprint matches the target fingerprint in the fingerprint syntax tree.

[0133] It should be noted that the fingerprint syntax tree mentioned in the embodiment is a sub-model of the fingerprint identification model described in the method. Figure 1 to Figure 2 In the embodiment, the fingerprint rule is determined preferentially, and then the fingerprint is determined.

[0134] When the to-be-identified fingerprint rule satisfies the target fingerprint rule, the to-be-identified fingerprint rule satisfying the target fingerprint rule and the to-be-identified fingerprint corresponding to the to-be-identified fingerprint rule are input into the fingerprint syntax tree, so that the to-be-identified fingerprint is matched by the fingerprint syntax tree, and it is determined whether the to-be-identified fingerprint matches the target fingerprint in the fingerprint syntax tree.

[0135] In one embodiment, based on the matching result of step 402, that is, the result of AC automaton matching, the syntax tree matching is performed to determine whether the corresponding to-be-identified fingerprint hits the target fingerprint. Since the AC automaton mainly matches the fingerprint rule, and the fingerprint rule is associated with the fingerprint, the corresponding target fingerprint is also associated with the target fingerprint rule. Therefore, when the to-be-identified fingerprint satisfying the target fingerprint rule is determined, only after the syntax tree is determined, it is determined whether the syntax tree exists the to-be-identified fingerprint, so as to determine whether the to-be-identified fingerprint matches the target fingerprint in the fingerprint syntax tree.

[0136] 404、If one of the to-be-identified fingerprints does not satisfy the target fingerprint rule, it is determined that the to-be-identified fingerprint is an abnormal fingerprint, and the next to-be-identified fingerprint rule is matched until all to-be-identified fingerprint rules are matched.

[0137] When one of the to-be-identified fingerprint rules does not satisfy the target fingerprint rule, the to-be-identified fingerprint corresponding to the to-be-identified fingerprint rule is an abnormal fingerprint. At the same time, since all to-be-identified fingerprints and to-be-identified fingerprint rules need to be determined, the next to-be-identified fingerprint rule needs to be matched until all to-be-identified fingerprint rules are matched. Specifically, it can be understood that after step 404 is executed, step 403 is executed again.

[0138] In one specific embodiment, the abnormal fingerprint can be understood as a fingerprint that cannot read an HTML page or other text, and cannot achieve the user's related operation requirements. Correspondingly, the abnormal fingerprint can also refer to a fingerprint that does not match the fingerprint in the fingerprint library, that is, it can also refer to a fingerprint that is not recorded in the fingerprint library, and the abnormal fingerprint is not limited here. For convenience of description, the abnormal fingerprint will not be described hereinafter.

[0139] 405、If the to-be-identified fingerprint is a target fingerprint satisfying the target fingerprint rule, the target fingerprint is matched according to the regular matching model to determine the fingerprint version type of the target fingerprint.

[0140] Based on step 403, when the to-be-identified fingerprint is a target fingerprint satisfying the target fingerprint rule, the target fingerprint can be matched using the regular matching model to determine the fingerprint version type of the target fingerprint.

[0141] In one specific embodiment, the regular matching is performed on the fingerprint matched in the syntax tree matching, so as to determine the fingerprint version type of the fingerprint. Because all regular matching is performed for version determination, the number of regular matching can be reduced to less than ten by using the already matched fingerprint, which greatly improves the matching efficiency.

[0142] Based on the above embodiment, the fingerprint library not only has fingerprints but also version information, such as Figure 1 The fingerprint mentioned in step 101 has a corresponding version Version,

[0143] Version: [{“v1”:“v1”},{“v2”:“V2”}]

[0144] Therefore, after determining the corresponding fingerprint name, the regular matching can be used to match the related fingerprint version.

[0145] When the fingerprint version is determined, a matching result can be generated. Specifically, the fingerprint matched by the syntax tree (i.e., the target fingerprint) and the fingerprint version matched by the regular expression are output. In one embodiment, the fingerprint name and the fingerprint version can be output. It should be understood that the output content is not limited herein and will not be described in detail.

[0146] 406. If the to-be-identified fingerprint does not match the target fingerprint, it is determined that the to-be-identified fingerprint is an abnormal fingerprint, and the next to-be-identified fingerprint is matched until all to-be-identified fingerprints are matched.

[0147] Based on step 403, when the to-be-identified fingerprint does not match the target fingerprint, it can be determined that the to-be-identified fingerprint is an abnormal fingerprint. For a detailed description of the abnormal fingerprint, please refer to step 404, which will not be described in detail here.

[0148] Since all to-be-identified fingerprints need to be determined, the next to-be-identified fingerprint needs to be matched until all to-be-identified fingerprints are matched. Specifically, it can be understood that after step 406 is executed, step 403 is executed again.

[0149] It should be understood that although each step in the flowchart involved in each of the above embodiments is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise stated herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, as described above, at least part of the steps in the flowchart involved in each of the above embodiments can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0150] If the scheme involves sensitive information (such as user information, enterprise information), it should be noted that the collection, use and processing of sensitive information need to comply with relevant national and regional laws, regulations and standards, and need to be permitted or agreed by the corresponding subject (such as users or enterprises, etc.).

[0151] Please refer to Figure 7 , Figure 7 A structure schematic diagram of a training system of a fingerprint identification model disclosed in an embodiment of the present application.

[0152] The acquisition unit 701 is configured to acquire a fingerprint library, wherein the fingerprint library includes to-be-matched fingerprints and to-be-matched fingerprint rules corresponding to the to-be-matched fingerprints, the to-be-matched fingerprint rules include to-be-matched rule strings and target rule relationships between different rule strings.

[0153] The generating unit 702 is configured to generate a fingerprint syntax tree according to the to-be-matched fingerprint and the target rule relationship, and construct a string matching model according to the to-be-matched rule string and the to-be-matched fingerprint.

[0154] The generating unit 702 is further configured to generate a fingerprint recognition model according to the fingerprint syntax tree and the string matching model, so as to recognize the fingerprint according to the fingerprint recognition model.

[0155] Exemplarily, the training system of the fingerprint recognition model further includes a ranking unit 703.

[0156] The ranking unit 703 is configured to re-rank the rule string with the rule relationship, so as to form a rule string with the target rule relationship.

[0157] The ranking unit 703 is further configured to rank the rule string and the to-be-matched fingerprint according to the target rule relationship, so as to form the fingerprint syntax tree.

[0158] Exemplarily, the training system of the fingerprint recognition model further includes a constructing unit 704 and a matching unit 705.

[0159] The constructing unit 704 is configured to construct an ordered tree by taking the to-be-matched rule string corresponding to the to-be-matched fingerprint as a node based on different to-be-matched fingerprints.

[0160] The matching unit 705 is configured to match the to-be-matched fingerprint and the to-be-matched rule string in the ordered tree, and obtain matching success information.

[0161] The constructing unit 704 is further configured to construct a string matching model according to the matching success information.

[0162] Please refer to Figure 8 , Figure 8 FIG. 1 is a structural schematic diagram of a fingerprint recognition system disclosed by an embodiment of the present application.

[0163] The obtaining unit 801 is configured to obtain a to-be-recognized sample, wherein the to-be-recognized sample includes a to-be-recognized fingerprint and a to-be-recognized fingerprint rule corresponding to the to-be-recognized fingerprint.

[0164] The input unit 802 is configured to input the to-be-recognized fingerprint and the to-be-recognized fingerprint rule to a pre-trained fingerprint recognition model, so as to determine whether the to-be-recognized fingerprint is a target fingerprint satisfying a target fingerprint rule according to the fingerprint recognition model; wherein the fingerprint recognition model is generated by the training method of the first aspect.

[0165] Exemplarily, the fingerprint recognition model includes a string matching model and a fingerprint syntax tree, and the fingerprint recognition system includes:

[0166] The input unit 802 is specifically configured to input all to-be-identified fingerprint rules to the string matching model, and match all to-be-identified fingerprint rules according to the string matching model, to determine to-be-identified fingerprint rules that meet the target fingerprint rule.

[0167] The input unit 802 is further configured to, when the to-be-identified fingerprint rule meets the target fingerprint rule, input the to-be-identified fingerprint rule that meets the target fingerprint rule and the to-be-identified fingerprint to the fingerprint syntax tree, and match the to-be-identified fingerprint according to the fingerprint syntax tree, to determine whether the to-be-identified fingerprint matches the target fingerprint in the fingerprint syntax tree.

[0168] Exemplarily, the fingerprint identification system further includes a determination unit 803.

[0169] The determination unit 803 is configured to, when one of the to-be-identified fingerprint rules does not meet the target fingerprint rule, determine that the to-be-identified fingerprint is an abnormal fingerprint, and match the next to-be-identified fingerprint rule until all to-be-identified fingerprint rules are matched.

[0170] Exemplarily, the fingerprint identification system includes:

[0171] The determination unit 803 is further configured to, when the to-be-identified fingerprint does not match the target fingerprint, determine that the to-be-identified fingerprint is an abnormal fingerprint, and match the next to-be-identified fingerprint until all to-be-identified fingerprints are matched.

[0172] Exemplarily, the fingerprint identification system further includes a matching unit 804.

[0173] The matching unit 804 is configured to, when the to-be-identified fingerprint is a target fingerprint that meets the target fingerprint rule, match the target fingerprint according to the regular matching model, to determine a fingerprint version type of the target fingerprint.

[0174] Please refer to Figure 9 The structure diagram of the training device of the fingerprint identification model disclosed in the embodiments of the present application includes:

[0175] The central processing unit 901, the memory 905, the input and output interface 904, the wired or wireless network interface 903 and the power supply 902;

[0176] The memory 905 is a transitory storage memory or a persistent storage memory;

[0177] The central processing unit 901 is configured to communicate with the memory 905, and execute instruction operations in the memory 905 to perform the foregoing Figure 1 Or Figure 2 The training method of the fingerprint identification model in the embodiments.

[0178] Please refer to Figure 10A structural schematic diagram of a fingerprint identification device disclosed by an embodiment of the present application comprises:

[0179] The central processor 1001, the memory 1005, the input and output interface 1004, the wired or wireless network interface 1003 and the power supply 1002;

[0180] The memory 1005 is a transitory storage memory or a persistent storage memory;

[0181] The central processor 1001 is configured to communicate with the memory 1005 and execute the instruction operation in the memory 1005 to perform the aforementioned Figure 3 or Figure 4 The fingerprint identification method in the illustrated embodiment.

[0182] The embodiment of the present application further provides a chip system, characterized in that the chip system comprises at least one processor and a communication interface, the communication interface and the at least one processor are interconnected through a line, and the at least one processor is used to run a computer program or instruction to execute the method in any one of the aforementioned Figure 1 to Figure 4 embodiments.

[0183] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0184] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented by other ways. For example, the device embodiments described above are merely schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division mode, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0185] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0186] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0187] When the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application, essentially or in the form of a contribution to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, read-only memory), a random access memory (RAM, random access memory), a magnetic disk or an optical disk, and various media that can store program codes.

Claims

1. A method for training a fingerprint recognition model, characterized in that, The method includes: Obtain a fingerprint database, wherein the fingerprint database includes fingerprints to be matched and matching fingerprint rules corresponding to the fingerprints to be matched, the matching fingerprint rules include rule strings to be matched and rule relationships between different rule strings; the fingerprints to be matched include website fingerprints and / or device fingerprints; A fingerprint syntax tree is generated based on the fingerprint to be matched and the rule relationship, and a string matching model is constructed based on the rule string to be matched and the fingerprint to be matched; the process of generating the fingerprint syntax tree is carried out using the scheduling field algorithm, and the process of constructing the string matching model is carried out using the AC automaton. The fingerprint recognition model is generated based on the fingerprint syntax tree and the string matching model, so as to identify the fingerprint based on the fingerprint recognition model.

2. The training method for the fingerprint recognition model according to claim 1, characterized in that, The step of generating a fingerprint syntax tree based on the fingerprint to be matched and the rule relationship includes: The rule strings with the aforementioned rule relationships are rearranged to form the rule strings with the target rule relationships; The rule string and the fingerprint to be matched are arranged according to the target rule relationship to form the fingerprint syntax tree.

3. The fingerprint recognition method according to claim 1, characterized in that, The step of constructing a string matching model based on the rule string to be matched and the fingerprint to be matched includes: Based on the different fingerprints to be matched, the rule strings to be matched corresponding to the fingerprints to be matched are used as nodes to construct an ordered tree; The fingerprints to be matched and the rule strings to be matched in the ordered tree are matched to obtain a successful match message; The string matching model is constructed based on the successful matching information.

4. A fingerprint recognition method, characterized in that, The method includes: Obtain a sample to be identified, wherein the sample to be identified includes a fingerprint to be identified and a fingerprint rule corresponding to the fingerprint to be identified; The fingerprint to be identified and the fingerprint rules to be identified are input into a pre-trained fingerprint recognition model to determine whether the fingerprint to be identified is a target fingerprint that satisfies the target fingerprint rules; wherein, the fingerprint recognition model is generated by the training method as described in any one of claims 1 to 3.

5. The fingerprint recognition method according to claim 4, characterized in that, The fingerprint recognition model includes a string matching model and a fingerprint syntax tree. The step of inputting the fingerprint to be recognized and the fingerprint rules to be recognized into the pre-trained fingerprint recognition model, and determining whether the fingerprint to be recognized is a target fingerprint that satisfies the target fingerprint rules based on the fingerprint recognition model, includes: All the fingerprint rules to be identified are input into the string matching model, and all the fingerprint rules to be identified are matched according to the string matching model to determine the fingerprint rules to be identified that satisfy the target fingerprint rules; If the fingerprint rule to be identified satisfies the target fingerprint rule, the fingerprint rule to be identified that satisfies the target fingerprint rule and the fingerprint to be identified are input into the fingerprint syntax tree, and the fingerprint to be identified is matched according to the fingerprint syntax tree to determine whether the fingerprint to be identified matches the target fingerprint in the fingerprint syntax tree.

6. The fingerprint recognition method according to claim 4, characterized in that, After determining whether the fingerprint to be identified is a target fingerprint that satisfies the target fingerprint rules based on the fingerprint recognition model, the method further includes: If the fingerprint to be identified is the target fingerprint that satisfies the target fingerprint rule, the target fingerprint is matched according to the regular matching model to determine the fingerprint version type of the target fingerprint.

7. A training system for a fingerprint recognition model, characterized in that, The system includes: An acquisition unit is configured to acquire a fingerprint database, wherein the fingerprint database includes fingerprints to be matched and fingerprint rules corresponding to the fingerprints to be matched, the fingerprint rules to be matched include rule strings to be matched and target rule relationships between different rule strings; the fingerprints to be matched include website fingerprints and / or device fingerprints; The generation unit is used to generate a fingerprint syntax tree based on the fingerprint to be matched and the target rule relationship, and to construct a string matching model based on the rule string to be matched and the fingerprint to be matched; the process of generating the fingerprint syntax tree is to use the scheduling field algorithm, and the process of constructing the string matching model is to use the AC automaton. The generation unit is further configured to generate the fingerprint recognition model based on the fingerprint syntax tree and the string matching model, so as to recognize the fingerprint based on the fingerprint recognition model.

8. A fingerprint recognition system, characterized in that, The system includes: An acquisition unit is used to acquire a sample to be identified, wherein the sample to be identified includes a fingerprint to be identified and a fingerprint rule corresponding to the fingerprint to be identified; An input unit is configured to input the fingerprint to be identified and the fingerprint rules to be identified into a pre-trained fingerprint recognition model, so as to determine whether the fingerprint to be identified is a target fingerprint that satisfies the target fingerprint rules according to the fingerprint recognition model; wherein, the fingerprint recognition model is generated by the training method as described in any one of claims 1 to 3.

9. A fingerprint recognition device, characterized in that, The device includes: Central processing unit, memory, input / output interfaces, wired or wireless network interfaces, and power supply; The memory is either a short-term storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute instructions in the memory to perform the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 6.

Images