WO2021000407A1

WO2021000407A1 - Character verification method and apparatus, and computer device and storage medium

Info

Publication number: WO2021000407A1
Application number: PCT/CN2019/103061
Authority: WO
Inventors: 李敏
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-07-03
Filing date: 2019-08-28
Publication date: 2021-01-07
Also published as: CN110351094B; CN110351094A

Abstract

A character verification method and apparatus, and a computer device and a storage medium. The method comprises: receiving a verification request sent by a target terminal (S1100); randomly extracting, according to the verification request, a verification statement from a pre-set statement database, wherein the verification statement is used for verifying a prompt (S1200); identifying the meaning of the verification statement, and generating verification characters according to a verification mode represented by the meaning of the verification statement, wherein at least one character in the verification characters has a mapping relationship with the verification mode (S1300); and sending the verification statement and the verification characters to the target terminal (S1400). A server end sends a verification statement and verification characters to a target terminal at the same time during verification, and a user can select a correct character from the verification characters to complete verification only after correctly understanding a semantic prompt of the verification statement. Therefore, only after the meaning of the verification statement is understood, can selective verification be performed on the verification characters by means of image recognition technology, and the vulnerability of a direct-input-type verification mode being easily cracked by means of the image recognition technology is avoided, such that verification security is improved, and network resources can be securely used.

Description

Character verification method, device, computer equipment and storage medium

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on July 3, 2019, the application number is 201910601303.1, and the invention title is "character verification method, device, computer equipment and storage medium", the entire content of which is incorporated by reference In this application.

Technical field

The embodiments of the present application relate to the field of data security, in particular, a character verification method, device, computer equipment, and storage medium.

Background technique

With the development of science and technology, the advent of the information age has brought us a lot of convenience, and at the same time, it has also brought a lot of trouble to people's lives. For example, when buying tickets online through the Internet, illegal vendors often use the development of applications to quickly swipe the tickets, and then resell them at high prices to obtain huge profits, but users who really need to buy cannot purchase through the Internet interface, and in real life, similar The snatching of Internet resources occurs in various fields, and it is difficult to stop the behavior of quickly swiping tickets and receiving commissions through the application. In order to limit the occurrence of the above behaviors, information verification applications were born.

The applicant realizes that in the prior art, verification codes are usually used for verification. When performing a verification operation, the server first obtains the verification code from the server, and then receives the verification information input by the user according to the verification code. The client sends the collected user information to the server, and the server determines whether the verification is passed by comparing whether the verification code is consistent with the text in the verification information.

The inventor of this application found in research that the verification code technology in the prior art simply sets the verification code on the background image for display, and the verification code can be identified without obstacles through the image recognition technology, and then the verification code is directly identified The verification code is sent to the server for verification without manual input. Therefore, the verification code in the prior art is easy to be identified, the verification security level is low, and it is impossible to truly protect network resources from being used safely.

Summary of the invention

The embodiment of the application provides a verification method that prompts the verification result through a verification sentence during verification. Only when the user understands the meaning of the verification sentence, can the user select the correct verification result character verification method, device, computer equipment and storage medium.

In order to solve the above technical problems, a technical solution adopted in the embodiment created by this application is to provide a character verification method, including:

Receive a verification request sent by the target terminal;

Randomly extract verification sentences from a preset sentence database according to the verification request, wherein the verification sentences are used for verification prompts;

Identifying the semantics of the verification sentence and generating verification characters according to a verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode;

Sending the verification sentence and the verification character to the target terminal.

To solve the above technical problems, an embodiment of the present application also provides a character verification device, including:

The acquisition module is used to receive the verification request sent by the target terminal;

A processing module, configured to randomly extract verification sentences from a preset sentence database according to the verification request, wherein the verification sentences are used for verification prompts;

A recognition module for recognizing the semantics of the verification sentence and generating verification characters according to the verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode;

The execution module is configured to send the verification sentence and the verification character to the target terminal.

In order to solve the above technical problems, an embodiment of the present application further provides a computer device including a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the The processor executes the steps of the character verification method, wherein the steps of the character verification method include:

Receive a verification request sent by the target terminal;

In order to solve the above technical problems, the embodiments of the present application also provide a non-volatile storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, one or more processing The device executes the steps of the character verification method described above, wherein the steps of the character verification method include:

Receive a verification request sent by the target terminal;

In this application, the server side simultaneously sends the verification sentence and verification character to the target terminal during verification. Only after the user correctly understands the semantic prompt of the verification sentence, can the user select the correct character from the verification character to complete the verification. Therefore, the image recognition technology can only select and verify the verification characters after understanding the semantics of the verification sentence, avoiding the vulnerability of the direct input verification method that is easy to be cracked by the image recognition technology, improving the security of verification, and protecting the network. Resources can be used safely.

Description of the drawings

Figure 1 is a schematic diagram of the basic flow of a character verification method according to an embodiment of this application;

FIG. 2 is a schematic diagram of the process of generating verification characters according to semantic recognition according to an embodiment of the application;

FIG. 3 is a schematic diagram of the process of deforming verification characters according to verification sentences according to an embodiment of the application; FIG.

FIG. 4 is a schematic diagram of a process of generating a verification image through image conversion according to an embodiment of the application;

FIG. 5 is a schematic diagram of a flow chart of performing display verification on a repeatedly verified target terminal according to an embodiment of the application;

FIG. 6 is a schematic diagram of the process of judging the verification result through target characters and semantic characters according to an embodiment of the application;

FIG. 7 is a schematic diagram of a process for judging whether a target character is consistent with a semantic character according to an embodiment of the application;

FIG. 8 is a schematic diagram of the basic structure of a character verification device according to an embodiment of the application;

Fig. 9 is a block diagram of the basic structure of a computer device according to an embodiment of the application.

Detailed ways

In order to enable those skilled in the art to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

Please refer to FIG. 1 for details. FIG. 1 is a schematic diagram of the basic flow of the character verification method in this embodiment.

As shown in Figure 1, a character verification method includes:

S1100. Receive a verification request sent by the target terminal.

When the user operates the target terminal to enter the application program or starts a certain functional module of the application program, the user instruction will trigger the corresponding verification function. For example, user actions such as application login, online purchase of various tickets, or e-commerce platform submission of orders, will trigger the user verification function.

After the verification function is triggered, the target terminal sends a verification request to the corresponding server, requesting to obtain verification data for verification.

S1200. Randomly extract verification sentences from a preset sentence database according to the verification request, where the verification sentences are used for verification prompts;

In this embodiment, the information sent by the server to the target terminal for verification includes a background image, a verification sentence and a verification character, wherein the verification sentence and the verification character are set on the background image.

A sentence database is set on the server side, and multiple verification sentences are preset in the sentence database. The verification sentences are extracted from the sentence database by random extraction or sequential extraction.

The verification sentence is a text field that prompts the user to select the correct characters for verification. For example, the verification sentence is (not limited to): Please select the text next to the word "女" in the following text, please select the font size of the following text is larger than normal For the font size, please select the text with a distorted font in the following text or the words that express "joy emotion" in the following text and other prompt sentences. The function of the verification sentence is to prompt the user to select the correct character for verification. Only after the user understands the verification sentence, can the user select the correct character in the verification character for verification. Therefore, as long as there is a suggestive meaning and verification When the characters have a mapping relationship with the suggestive meaning, the text field can be used as a verification sentence.

S1300. Identify the semantics of the verification sentence and generate verification characters according to the verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode;

After the verification sentence is obtained by random extraction, the verification mode represented by the text field recorded in the verification sentence is identified. The recognition method is to train a neural network model that can recognize semantics according to deep learning, and then use the neural network model to recognize the semantics of the verification sentence.

In this embodiment, the verification method of the verification sentence representation can be divided into: selecting characters based on pronunciation, selecting characters based on structure, selecting characters based on glyph shape, or selecting characters based on meaning. For example, the verification method of the verification sentence characterization of "select the character whose initials are y in the following characters" is to select the character according to the pronunciation; the verification method of the verification sentence characterization of the verification sentence characterization of "select the following characters with the word "女" is the basis Structure selection character; the verification method of the verification sentence characterization of "please select the distortion in the following text" is to select the character according to the shape of the character; .

According to the verification mode represented by the verification sentence, the corresponding verification character is extracted from the preset character database, and at least one character in the extracted verification character is mapped to the verification sentence. For example, the verification sentence is: "Select the text next to the word "女" in the following text", and the verification characters are: "女", "高", "rate", and "祯".

The verification character is a text field composed of multiple text characters that carries the correct character. Each text character can be independent of each other and does not form a specific word, but the composition method of the verification character is not limited to this. According to different application scenarios, some In an embodiment, the verification character consists of multiple words. The verification characters include regular characters and characters that have a mapping relationship with the verification sentence. Among them, the characters that have a mapping relationship with the verification sentence are correct characters, and the user can complete the character verification by selecting the above-mentioned characters; otherwise, the character verification cannot be completed.

S1400. Send the verification sentence and the verification character to the target terminal.

The generated verification sentence and verification characters are sent to the target terminal, so that the target terminal performs character verification.

In some embodiments, the verification data sent by the server to the target terminal further includes: a background image, the verification characters and the verification sentence are set on the background image, the verification picture is synthesized, and then the verification picture is sent to the target terminal for verification.

In the foregoing embodiment, the server sends the verification sentence and the verification character to the target terminal at the same time during verification, and the user can select the correct character from the verification character to complete the verification only after correctly understanding the semantic prompt of the verification sentence. Therefore, the image recognition technology can only select and verify the verification characters after understanding the semantics of the verification sentence, avoiding the vulnerability of the direct input verification method that is easy to be cracked by the image recognition technology, improving the security of verification, and protecting the network. Resources can be used safely.

Further, referring to FIG. 2, in an optional implementation manner, the step S1300 shown in FIG. 1 includes:

S1311: Call the verification sentence and read the text field in the verification sentence;

The verification sentence is a text field with specific semantics composed of characters, and the text field of the verification sentence is first read when performing character verification on the verification sentence.

S1312. Convert the text field into an array matrix according to a preset conversion method.

After reading the text field, the text field is converted into an array matrix through the matlab software application software. Among them, each word or word constituting the text field is mapped and replaced with an element of the array matrix, and the arrangement order of the elements is consistent with the arrangement order of the text field.

S1313. Input the data matrix into a preset semantic recognition model, where the semantic recognition model is a neural network model pre-trained to a convergent state for semantic recognition of text;

The semantics of the verification sentence is detected through the semantic recognition model. The semantic recognition model is a neural network model trained to a convergent state for classifying the semantics expressed by the verification sentence. Specifically, through matlab software, the verification sentence to be detected is converted into an array matrix that can be recognized by the neural network model, and then the array matrix is input into the semantic recognition model to obtain the semantics of the expression verification sentence output by the model The classification results.

The semantic recognition model in this embodiment can be a convolutional neural network model (CNN) that has been trained to a convergent state, but is not limited to this, the semantic recognition model can also be: deep neural network model (DNN), recurrent neural network model (RNN) or a deformed model of the above three network models.

When training the initial neural network model as a semantic recognition model, it collects a large number of converted array matrices of verification sentences as training samples. After manually reading the original verification sentences of the training samples, each training sample is calibrated (calibration refers to The semantics actually expressed by the training samples). Then input the training sample into the initial neural network model, and obtain the classification result of the model output (the classification result is the semantic classification of the training sample obtained by the model), and calculate the difference between the classification result and the calibration result through the loss function of the neural network model Compare the calculated result with the set distance threshold (for example, Euclidean distance, Mahalanobis distance, or cosine distance, etc.). If the calculated result is less than or equal to the distance threshold (for example, 0.05), pass the verification and continue. For the training of the next training sample, if the calculation result is greater than the distance threshold, the difference between the two is calculated through the loss function, and the weights in the neural network model are corrected through back propagation, so that the neural network model can improve the training sample The weight of elements corresponding to words that accurately express the semantics of the text can be used to increase the accuracy of judgment. After cyclically executing the above scheme and training with a large number of training samples, the neural network model obtained by training has a semantic judgment accuracy rate greater than a certain value, for example, 97%, and the neural network model is trained to a convergent state. The neural network trained to convergence is the semantic recognition model.

The semantic recognition model trained to the convergent state can accurately extract the semantics represented by the array matrix.

S1314. Extract the verification character from a preset character database according to the verification mode represented by the semantic classification result output by the semantic recognition model.

According to the semantic classification result output by the semantic recognition model, the verification character of the first character length is extracted from the character database.

For example, when the verification sentence is: Please select the fourth and sixth characters in the following text. At this time, the semantic classification result output by the semantic recognition model is: "Verify according to the character number, and the largest character needs to be the sixth digit". Then, the verification characters that are greater than or equal to six characters are extracted from the verification character library.

When the verification sentence is: Please select the characters next to the word "女" in the following text, the semantic classification result output by the semantic recognition model is: "Verify according to the radical, verify that the radical is female". Then at least one character is extracted from the character database as a verification character next to the female character.

When the verification sentence is: Please select the sentence of characters with bold font in the following text, the semantic classification result output by the semantic recognition model is: "Verify according to the font shape, and the verification font is bold".

Semantic recognition of the verification sentence through the neural network model enables the terminal to generate verification characters corresponding to the semantic recognition results, which improves the efficiency and accuracy of verification character generation, thereby increasing the efficiency of verification.

In some embodiments, the verification method of the verification sentence characterization is to perform verification according to the shape of the font. Therefore, when the verification character is generated, some characters need to be deformed according to the semantics of the verification sentence.

Further, referring to FIG. 3, in an optional implementation manner, after step S1314 shown in FIG. 2, the method further includes:

S1321, filter at least one deformed character among the verification characters;

When the verification mode represented by the verification sentence is to perform verification by selecting characters from the font, after the characters constituting the verification characters are obtained, at least one deformed character is selected from the verification characters. In this embodiment, the number of characters selected as deformed characters in the verification character is not limited to one. According to different application scenarios, in some embodiments, the number of deformed characters can be greater until all characters in the verification character are equal. It is a deformed character.

S1322, performing deformation processing on the at least one deformed character according to the deformation type represented by the verification mode to generate semantic characters, wherein the semantic characters have a mapping relationship with the verification sentence;

According to the semantic classification result determined by the semantic recognition model, at least one deformed character is deformed by the represented deformation type to generate a semantic character.

The deformation types in this embodiment include (not limited to): enlarging, reducing, bolding, slanting, rotating, twisting, or changing colors of characters.

Characters generated after deformation processing are defined as semantic characters, and there is a mapping relationship between semantic characters and verification sentences, that is, phonetic characters are the standard answers for character verification.

S1323. Replace the deformed character in the verification character with the semantic character.

The deformed character in the verification character is replaced with the generated semantic character, that is, the correct answer of the verification is written into the verification character, and the deformation processing of the verification character is completed. At the same time, the semantic characters are stored, and the semantic characters are used as the correct answer of the verification sentence for verification and comparison.

By using character deformation processing, the result of character verification can be made more prominent, which is conducive to the user's choice. At the same time, the character shape change is not easily recognized by the image recognition technology, which improves the security of character verification.

In some embodiments, the verification characters and verification sentences need to be displayed in a picture format. Therefore, it is necessary to image the verification characters and verification sentences.

Further, referring to FIG. 4, in an optional implementation manner, the step S1400 shown in FIG. 1 includes:

S1411, filter background images in a preset image database;

In this embodiment, the server is provided with an image database, and various pictures are pre-stored in the image database, and the pictures stored in the image database can be used for character verification.

After the semantic character replaces the deformed character, a picture is selected as a background image through random screening in the image database. However, the screening method is not limited to this. According to different application scenarios, in some embodiments, when screening background images, it is necessary to screen background images with a color difference equal to 2 or 3 between the font colors of the verification characters to increase verification. The degree of fusion between characters and background images makes it difficult for image recognition technology to recognize and verify characters, which increases the difficulty of machine recognition.

S1412, overlaying the verification sentence and the verification character on the background image to generate a verification image, wherein the verification character includes the semantic character;

Set the verification sentence and the verification character on the background image, and set the verification sentence and the verification character on the top when setting the image level, so that the verification sentence and the verification character are overlaid on the background image. After the setting is completed, a verification image with verification sentences and verification characters is generated.

S1413. Send the verification image to the target terminal.

After the verification image is generated, the verification image is sent to the target terminal so that the target terminal can use the verification image for character verification.

Setting the verification character and the verification sentence in the background image can increase the degree of confusion between the verification sentence and the verification character and the background, and increase the difficulty of recognition.

In some embodiments, in order to prevent the same terminal from passing virtual verification, the verification system is quickly swiped. The server records the verification records of each target terminal, and verifies the virtual swiping behavior based on the verification records. Please refer to FIG. 5. FIG. 5 is a schematic diagram of the display verification process of the repeatedly verified target terminal in this embodiment.

Further, referring to FIG. 5, in an optional implementation manner, after step S1413 shown in FIG. 4, the method further includes:

S1421, search for the verification record of the target terminal in the historical verification list;

In this embodiment, the record of each target terminal's verification on the server side is recorded in the history verification list, and the recording method is: recording the identity information of the target terminal, for example, the IP address or Mac address of the target terminal. Then, each verification request of the target terminal is recorded under the identity information of the corresponding target terminal in the history verification list. Therefore, as long as the identity information of the target terminal in the verification request is obtained, the corresponding verification record can be found in the historical verification list.

S1422, when the verification record of the target terminal meets the preset first verification condition, call the screenshot screen of the verification image in the verification page to the target terminal;

The verification record records the verification frequency of the target terminal within a set time period, for example, the target terminal is verified 100 times in one hour. The length of the time period for statistical verification frequency in the verification record can be adjusted according to different specific application scenarios.

When it is detected whether the frequency of the target terminal in the set time period meets the set first verification condition, where the first verification condition refers to the value of the verification frequency in the set time period. For example, within a set hour, the number of verifications performed by the target terminal shall not exceed 10 times, but the set frequency value of the first verification condition is not limited to this, and it can be adapted according to the specific application scenarios. Increase or decrease in sex.

When the verification frequency of the target terminal in the set period of time is greater than or equal to the verification frequency set by the first verification condition, the server sends a call request to the target terminal, and calls the target terminal to display a screenshot of the verification image on the verification page. Identify whether the user uses virtual authentication to verify.

When the target terminal displays the verification image, it needs to store the verification page including the verification image in the frame buffer memory, that is, the frame buffer memory is a direct image of the screen displayed on the screen, which is also called a bit map. That is, the data is displayed.

Since the verification image has a set area in the bitmap, according to the information of the set area, the data area representing the content of the verification area is extracted from the bitmap to generate a local bitmap, that is, the target data that represents the display content of the verification image .

Finally, the target data is converted into a conventional picture format, such as (not limited to) JPG, PNG, or TIF, etc., to generate a verification image.

In some embodiments, when the verification image cannot be obtained in the frame buffer memory, it indicates that the verification method is virtual verification.

When the verification frequency of the target terminal within the set time period is less than the verification frequency set by the first verification condition, there is no need to call the screenshot screen to the target terminal.

S1423. Input the screenshot and the verification image into a preset verification judgment model, where the verification judgment model is a neural network model that is pre-trained to a convergence state and is used to judge image similarity;

The screenshot and the verification image are input into a preset verification judgment model, where the verification judgment model is a neural network model that is pre-trained to a convergence state for judging image similarity.

In this embodiment, the verification judgment model can be a convolutional neural network model (CNN) that has been trained to a convergent state, but is not limited to this. The verification judgment model can also be: deep neural network model (DNN), recurrent neural network model (RNN) or a deformed model of the above three network models.

The verification judgment model is a neural network model with a convergent shape, so it can accurately and quickly determine whether the verification images are the same or different.

S1424: Confirm whether the target terminal is a virtual verification according to the judgment result output by the verification judgment model.

The judgment result output by the verification judgment model contains a judgment conclusion that the verification judgment model judges whether the screenshot screen and the verification image are the same. When the screenshot is the same as the verification image, the verification of the target terminal is confirmed to be non-virtual verification; otherwise, it is determined that the target terminal performs virtual verification, and it is forbidden to respond to the request of the target terminal to access the server.

The server side judges whether the verification image and the screenshot are the same, which can prevent the virtual verifier from writing image data to the frame buffer memory by deception to avoid detection, which improves the accuracy of verification and further ensures the security of network data Sex.

In some embodiments, the server obtains the target character selected by the user for verification uploaded by the target terminal, and judges the result of the verification based on whether the target character is consistent with the semantic character.

Further, referring to FIG. 6, in an optional implementation manner, after the step S1400 shown in FIG. 1, the method further includes:

S1431. Acquire verification information uploaded by the target terminal, where the verification information includes target characters for verification that the user screens among the verification characters according to the verification sentence;

After sending the verification sentence and the verification characters to the target terminal, the server waits to receive the verification information uploaded by the target terminal, where the verification information includes the target characters.

The target character is a character selected in the verification character according to the user's click instruction. Among them, the characters in the target character may have a mapping relationship with the verification sentence; there may be only individual characters or no characters that have a mapping relationship with the verification sentence. When the characters in the target character all have a mapping relationship with the verification sentence, the verification passes; otherwise, the verification fails.

S1432. Compare the target character with the semantic character to determine whether the target character is consistent with the semantic character.

The server compares the target character with the semantic character, and the comparison method is to calculate the Hamming distance or the Hamming distance between the target character and the semantic character. When the Hamming distance or Hamming distance between the target character and the semantic character is zero, the target character is consistent with the semantic character; otherwise, the target character is inconsistent with the semantic character.

S1433: When the target character is consistent with the semantic character, confirm that the target terminal passes the verification.

When the target character is consistent with the semantic character, the server confirms that the target terminal character verification has passed; otherwise, it confirms that the target terminal verification fails.

In some embodiments, the Hamming distance or Hamming distance between the target character and the semantic character is calculated to determine whether the target character is consistent with the phonetic character.

Further, referring to FIG. 7, in an optional implementation manner, the step S1432 shown in FIG. 6 includes:

S1441, calculate the Hamming distance between the target character and the semantic character;

Calculate the Hamming distance between the target character and the semantic character. The Hamming distance is the different number of bits corresponding to two fields (same length). For example, the Hamming distance between "toned" and "roses" is 3.

S1442. When the Hamming distance is zero, confirm that the target character is consistent with the semantic character; otherwise, confirm that the target character is inconsistent with the semantic character.

When the Hamming distance between the target character and the semantic character is zero, the target character is consistent with the semantic character; otherwise, the target character is inconsistent with the semantic character.

Through the Hamming distance, the target character and the semantic character can be quickly compared, which improves the efficiency of the comparison and further improves the verification efficiency on the server side.

In order to solve the above technical problems, an embodiment of the present application also provides a character verification device.

Please refer to FIG. 8 for details. FIG. 8 is a schematic diagram of the basic structure of the character verification device in this embodiment.

As shown in FIG. 8, a character verification device includes: an acquisition module 2100, a processing module 2200, an identification module 2300, and an execution module 2400. Among them, the acquisition module 2100 is used to receive a verification request sent by the target terminal; the processing module 2200 is used to extract verification sentences from a preset sentence database according to the verification request, wherein the verification sentences are used for verification prompts; the identification module 2300 is used to identify verification The sentence and the verification character are generated according to the verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode; the execution module 2400 is used to send the verification sentence and the verification character to the target terminal.

The character verification device simultaneously sends the verification sentence and verification character to the target terminal during verification, and the user can select the correct character from the verification character to complete verification only after correctly understanding the semantic prompt of the verification sentence. Therefore, the image recognition technology can only select and verify the verification characters after understanding the semantics of the verification sentence, avoiding the vulnerability of the direct input verification method that is easy to be cracked by the image recognition technology, improving the security of verification, and protecting the network. Resources can be used safely.

In some embodiments, the character verification apparatus further includes: a first acquisition submodule, a first processing submodule, a first recognition submodule, and a first execution submodule. Among them, the first obtaining submodule is used to obtain the text field of the verification sentence; the first processing submodule is used to convert the text field into an array matrix according to a preset conversion method; the first recognition submodule is used to input the data matrix into the preset In the semantic recognition model set, the semantic recognition model is a neural network model that is pre-trained to a convergent state for semantic recognition of text; the first execution sub-module is used to represent the verification method based on the semantic classification result output by the semantic recognition model , Extract verification characters from the preset character database.

In some embodiments, the verification method is font verification, and the character verification device further includes: a first screening submodule, a second processing submodule, and a second execution submodule. Wherein, the first screening submodule is used for screening at least one deformed character in the verification character; the second processing submodule is used for deforming at least one deformed character according to the deformation type represented by the verification mode to generate semantic characters, wherein the semantic characters The verification statement has a mapping relationship; the second execution sub-module is used to replace deformed characters in the verification characters with semantic characters.

In some embodiments, the character verification device further includes: a second screening submodule, a third processing submodule, and a third execution submodule. Among them, the second screening sub-module is used to screen the background image in the preset image database; the third processing sub-module is used to overwrite the verification sentence and the verification characters on the background image to generate a verification image, wherein the verification characters include semantic characters ; The third execution submodule is used to send the verification image to the target terminal.

In some embodiments, the character verification device further includes: a first search submodule, a fourth processing submodule, a first input submodule, and a fourth execution submodule. Among them, the first search submodule is used to search the verification record of the target terminal in the historical verification list; the fourth processing submodule is used to call the verification image to the target terminal when the verification record of the target terminal meets the preset first verification condition The screenshot screen in the verification page; the first input sub-module is used to input the screenshot screen and the verification image into the preset verification judgment model, where the verification judgment model is pre-trained to the convergence state, and is used to judge image similarity Neural network model; the fourth execution sub-module is used to confirm whether the target terminal is a virtual verification according to the judgment result output by the verification judgment model.

In some embodiments, the character verification device further includes: a second acquisition submodule, a fifth processing submodule, and a fifth execution submodule. Wherein, the second obtaining submodule is used to obtain the verification information uploaded by the target terminal, where the verification information includes the target characters for verification selected by the user according to the verification sentence among the verification characters; the fifth processing submodule is used to transfer the target characters It is compared with semantic characters to determine whether the target character is consistent with the semantic character; the fifth execution submodule is used to confirm that the target terminal passes the verification when the target character is consistent with the semantic character.

In some embodiments, the character verification device further includes: a first calculation submodule and a sixth execution submodule. Among them, the first calculation submodule is used to calculate the Hamming distance between the target character and the semantic character; the sixth execution submodule is used to confirm that the target character is consistent with the semantic character when the Hamming distance is zero; otherwise, confirm the target Characters are inconsistent with semantic characters.

In order to solve the above technical problems, the embodiments of the present application also provide computer equipment. Please refer to FIG. 9 for details. FIG. 9 is a block diagram of the basic structure of the computer device in this embodiment.

As shown in Figure 9, a schematic diagram of the internal structure of the computer equipment. The computer device includes a processor, a nonvolatile storage medium, a memory, and a network interface connected through a system bus. Wherein, the non-volatile storage medium of the computer device stores an operating system, a database, and computer-readable instructions. The database may store control information sequences. When the computer-readable instructions are executed by the processor, the processor can realize a A method of character verification. The processor of the computer equipment is used to provide calculation and control capabilities, and supports the operation of the entire computer equipment. A computer readable instruction may be stored in the memory of the computer device, and when the computer readable instruction is executed by the processor, the processor may execute a character verification method. The network interface of the computer device is used to connect and communicate with the terminal. Those skilled in the art can understand that the structure shown in FIG. 9 is only a block diagram of part of the structure related to the solution of the present application, and does not constitute a limitation on the computer equipment to which the solution of the present application is applied. The specific computer equipment may Including more or fewer parts than shown in the figure, or combining some parts, or having a different arrangement of parts.

In this embodiment, the processor is used to execute the specific functions of the acquisition module 2100, processing module 2200, identification module 2300, and execution module 2400 in FIG. 8, and the memory stores program codes and various data required to execute the above modules. The network interface is used for data transmission between user terminals or servers. The memory in this embodiment stores the program codes and data required to execute all the sub-modules in the human face image key point detection device, and the server can call the program codes and data of the server to execute the functions of all the sub-modules.

The computer device simultaneously sends the verification sentence and the verification character to the target terminal at the same time during verification, and the user can select the correct character from the verification character to complete the verification only after correctly understanding the semantic prompt of the verification sentence. Therefore, the image recognition technology can only select and verify the verification characters after understanding the semantics of the verification sentence, avoiding the vulnerability of the direct input verification method that is easy to be cracked by the image recognition technology, improving the security of verification, and protecting the network. Resources can be used safely.

The present application also provides a storage medium storing non-volatile computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the character verification method in any of the above embodiments. A step of.

A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing relevant hardware through a computer program. The computer program can be stored in a computer readable storage medium. When executed, it may include the processes of the above-mentioned method embodiments. Among them, the aforementioned storage medium may be a non-volatile storage medium such as a magnetic disk, an optical disc, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM), etc.

It should be understood that, although the various steps in the flowchart of the drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless explicitly stated in this article, the execution of these steps is not strictly limited in order, and they can be executed in other orders. Moreover, at least part of the steps in the flowchart of the drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the order of execution is also It is not necessarily performed sequentially, but may be performed alternately or alternately with other steps or at least a part of sub-steps or stages of other steps.

Claims

A character verification method, including:

Receive a verification request sent by the target terminal;

Randomly extract verification sentences from a preset sentence database according to the verification request, wherein the verification sentences are used for verification prompts;

Identifying the semantics of the verification sentence and generating verification characters according to a verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode;

Sending the verification sentence and the verification character to the target terminal.
The character verification method according to claim 1, wherein the recognizing the semantics of the verification sentence and generating verification characters according to the verification mode represented by the semantics of the verification sentence comprises:

Call the verification sentence and read the text field in the verification sentence;

Converting the text field into an array matrix according to a preset conversion method;

Inputting the data matrix into a preset semantic recognition model, where the semantic recognition model is a neural network model pre-trained to a convergent state for semantic recognition of text;

According to the verification mode represented by the semantic classification result output by the semantic recognition model, the verification character is extracted from a preset character database.
The character verification method according to claim 2, wherein the verification method is font verification, and the verification method characterized by the semantic classification result output by the semantic recognition model is extracted from a preset character database after the verification character , The method further includes:

Screening at least one deformed character among the verification characters;

Performing a deformation process on the at least one deformed character according to the deformation type represented by the verification mode to generate a semantic character, wherein the semantic character has a mapping relationship with the verification sentence;

Replace the deformed character in the verification character with the semantic character.
The character verification method according to claim 3, wherein the sending the verification sentence and the verification character to the target terminal comprises:

Filter background images in a preset image database;

Overlaying the verification sentence and the verification character on the background image to generate a verification image, wherein the verification character includes the semantic character;

Sending the verification image to the target terminal.
The character verification method according to claim 4, after the sending the verification image to the target terminal, the method further comprises:

Search for the verification record of the target terminal in the historical verification list;

When the verification record of the target terminal meets the preset first verification condition, calling the target terminal a screenshot of the verification image in the verification page;

Inputting the screenshot and the verification image into a preset verification judgment model, where the verification judgment model is a neural network model that is pre-trained to a convergent state and used to judge image similarity;

According to the judgment result output by the verification judgment model, it is confirmed whether the target terminal is a virtual verification.
The character verification method according to claim 3, after the verification sentence and the verification character are sent to the target terminal, the method further comprises:

Acquiring verification information uploaded by the target terminal, where the verification information includes target characters for verification that the user screened among the verification characters according to the verification sentence;

Comparing the target character with the semantic character to determine whether the target character is consistent with the semantic character;

When the target character is consistent with the semantic character, it is confirmed that the target terminal passes the verification.
The character verification method according to claim 6, wherein the comparing the target character with the semantic character to determine whether the target character is consistent with the semantic character comprises:

Calculating the Hamming distance between the target character and the semantic character;

When the Hamming distance is zero, it is confirmed that the target character is consistent with the semantic character; otherwise, it is confirmed that the target character is inconsistent with the semantic character.
A character verification device includes:

The acquisition module is used to receive the verification request sent by the target terminal;

A processing module, configured to randomly extract verification sentences from a preset sentence database according to the verification request, wherein the verification sentences are used for verification prompts;

A recognition module for recognizing the semantics of the verification sentence and generating verification characters according to the verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode;

The execution module is configured to send the verification sentence and the verification character to the target terminal.
A computer device, comprising a memory and a processor, the memory stores computer readable instructions, and when the computer readable instructions are executed by the processor, the processor executes the steps of the character verification method described above , Wherein the steps of the character verification method include:

Receive a verification request sent by the target terminal;

Randomly extract verification sentences from a preset sentence database according to the verification request, wherein the verification sentences are used for verification prompts;

Identifying the semantics of the verification sentence and generating verification characters according to a verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode;

Sending the verification sentence and the verification character to the target terminal.
8. The computer device according to claim 9, wherein the recognizing the semantics of the verification sentence and generating verification characters according to a verification method characterized by the semantics of the verification sentence comprises:

Call the verification sentence and read the text field in the verification sentence;

Converting the text field into an array matrix according to a preset conversion method;

Inputting the data matrix into a preset semantic recognition model, where the semantic recognition model is a neural network model pre-trained to a convergent state for semantic recognition of text;

According to the verification mode represented by the semantic classification result output by the semantic recognition model, the verification character is extracted from a preset character database.
The computer device according to claim 10, wherein the verification method is character shape verification, and the verification method characterized by the semantic classification result output by the semantic recognition model is extracted from a preset character database after the verification character is extracted, The method also includes:

Screening at least one deformed character among the verification characters;

Performing a deformation process on the at least one deformed character according to the deformation type represented by the verification mode to generate a semantic character, wherein the semantic character has a mapping relationship with the verification sentence;

Replace the deformed character in the verification character with the semantic character.
The computer device according to claim 11, wherein the sending the verification sentence and the verification character to the target terminal comprises:

Filter background images in a preset image database;

Overlaying the verification sentence and the verification character on the background image to generate a verification image, wherein the verification character includes the semantic character;

Sending the verification image to the target terminal.
The computer device according to claim 12, after the sending the verification image to the target terminal, the method further comprises:

Search for the verification record of the target terminal in the historical verification list;

When the verification record of the target terminal meets the preset first verification condition, calling the target terminal a screenshot of the verification image in the verification page;

Inputting the screenshot and the verification image into a preset verification judgment model, where the verification judgment model is a neural network model that is pre-trained to a convergent state and used to judge image similarity;

According to the judgment result output by the verification judgment model, it is confirmed whether the target terminal is a virtual verification.
The computer device according to claim 11, after the sending the verification sentence and the verification character to the target terminal, the method further comprises:

Acquiring verification information uploaded by the target terminal, where the verification information includes target characters for verification that the user screened among the verification characters according to the verification sentence;

Comparing the target character with the semantic character to determine whether the target character is consistent with the semantic character;

When the target character is consistent with the semantic character, it is confirmed that the target terminal passes the verification.
A storage medium storing non-volatile computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the character verification method described above, Wherein, the steps of the character verification method include:

Receive a verification request sent by the target terminal;

Randomly extract verification sentences from a preset sentence database according to the verification request, wherein the verification sentences are used for verification prompts;

Identifying the semantics of the verification sentence and generating verification characters according to a verification mode represented by the semantics of the verification sentence, wherein at least one character in the verification character has a mapping relationship with the verification mode;

Sending the verification sentence and the verification character to the target terminal.
The non-volatile computer-readable instruction storage medium according to claim 15, wherein the recognizing the semantics of the verification sentence and generating verification characters according to a verification method characterized by the semantics of the verification sentence comprises:

Call the verification sentence and read the text field in the verification sentence;

Converting the text field into an array matrix according to a preset conversion method;

Inputting the data matrix into a preset semantic recognition model, where the semantic recognition model is a neural network model pre-trained to a convergent state for semantic recognition of text;

According to the verification mode represented by the semantic classification result output by the semantic recognition model, the verification character is extracted from a preset character database.
The computer device according to claim 15, wherein the verification method is character shape verification, and the verification method characterized by the semantic classification result output by the semantic recognition model is extracted from a preset character database after the verification character is extracted, The method also includes:

Screening at least one deformed character among the verification characters;

Performing a deformation process on the at least one deformed character according to the deformation type represented by the verification mode to generate a semantic character, wherein the semantic character has a mapping relationship with the verification sentence;

Replace the deformed character in the verification character with the semantic character.
The computer device according to claim 17, wherein the sending the verification sentence and the verification character to the target terminal comprises:

Filter background images in a preset image database;

Overlaying the verification sentence and the verification character on the background image to generate a verification image, wherein the verification character includes the semantic character;

Sending the verification image to the target terminal.
The computer device according to claim 18, after the sending the verification image to the target terminal, the method further comprises:

Search for the verification record of the target terminal in the historical verification list;

When the verification record of the target terminal meets the preset first verification condition, calling the target terminal a screenshot of the verification image in the verification page;

Inputting the screenshot and the verification image into a preset verification judgment model, where the verification judgment model is a neural network model that is pre-trained to a convergent state and used to judge image similarity;

According to the judgment result output by the verification judgment model, it is confirmed whether the target terminal is a virtual verification.
The computer device according to claim 17, after the sending the verification sentence and the verification character to the target terminal, the method further comprises:

Acquiring verification information uploaded by the target terminal, where the verification information includes target characters for verification that are selected by the user according to the verification sentence among the verification characters;

Comparing the target character with the semantic character to determine whether the target character is consistent with the semantic character;

When the target character is consistent with the semantic character, it is confirmed that the target terminal passes the verification.