Access verification method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application provide an access verification method and device, electronic equipment and storage medium, and relate to the technical field of data verification. The method comprises: receiving prompt text randomly generated and sent by an access device, generating at least one first image conforming to the description of the prompt text and at least one second image not conforming to the description of the prompt text according to the prompt text; sending the first image and the second image to the access device; receiving a target image sent by the access device; and determining a verification result based on the first image and the target image. In this way, in the verification process, the first image and the second image that have never appeared are generated according to the randomly generated prompt text, the access verification is performed through the first image and the second image, the model for cracking the verification mechanism cannot learn the randomly generated image, the image verification code is prevented from being cracked due to image library leakage, and the verification security is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data verification, and particularly relates to an access verification method and device, electronic equipment and a storage medium. BACKGROUND

[0002] A verification code is a public automatic program to distinguish between computer operation and human operation. Among them, image verification is a new generation of Internet verification mechanism, which is completed through the form of clicking or sliding the image verification code.

[0003] In the related art, in order to prevent websites or software from being intruded, the verification image of the image verification code is usually set to be very blurred, and the verification part in the image verification code is deformed and rotated and complex interference is added. However, when the image library providing the image verification code is leaked, a model that can crack the verification mechanism can be trained through the image library to easily crack the image verification mechanism, thereby achieving the purpose of intrusion. Therefore, it is urgent to provide an access verification method to solve the above technical problems. SUMMARY

[0004] The main purpose of the embodiments of the present application is to provide an access verification method, device, electronic equipment and storage medium, which can avoid the image verification code being cracked by the model that cracks the verification mechanism, and improve the verification security.

[0005] To achieve the above purpose, a first aspect of the embodiments of the present application provides an access verification method, which comprises:

[0006] receiving prompt text randomly generated and sent by an access device, generating at least one first image conforming to the description of the prompt text and at least one second image not conforming to the description of the prompt text according to the prompt text;

[0007] sending the first image and the second image to the access device, so that the prompt text is displayed in a first area in a verification control of the access device, and each of the first image and each of the second image is displayed in a second area in the verification control;

[0008] receiving a target image sent by the access device, the target image being an image determined by the access device according to an image selection operation on the second area;

[0009] determining a verification result based on the first image and the target image.

[0010] In some embodiments, the at least one first image conforming to the description of the prompt text and the at least one second image not conforming to the description of the prompt text are generated according to the prompt text, comprising:

[0011] determine a first text vector corresponding to the prompt text;

[0012] change a target word in the prompt text that meets a change condition to obtain a modified text;

[0013] determine a second text vector corresponding to the modified text;

[0014] generate at least one first image that meets the description of the prompt text based on the first text vector;

[0015] generate at least one second image that does not meet the description of the prompt text based on the second text vector.

[0016] In some embodiments, the changing a target word in the prompt text that meets a change condition to obtain a modified text comprises:

[0017] obtain a plurality of candidate words that constitute the prompt text;

[0018] determine a word type corresponding to each of the candidate words;

[0019] determine a candidate word in the plurality of candidate words as a target word if the word type of the candidate word is a preset word type;

[0020] change at least one target word in the prompt text to a word of the same word type to obtain a modified text.

[0021] In some embodiments, the determining a first text vector corresponding to the prompt text comprises:

[0022] perform word segmentation processing on the prompt text to obtain a plurality of first words;

[0023] encode each of the first words to obtain a corresponding first word vector;

[0024] combine the plurality of first word vectors into a first text vector;

[0025] The determining a second text vector corresponding to the modified text comprises:

[0026] perform word segmentation processing on the modified text to obtain a plurality of second words;

[0027] encode each of the second words to obtain a corresponding second word vector;

[0028] combine the plurality of second word vectors into a second text vector.

[0029] In some embodiments, the determining a verification result based on the first image and the target image comprises:

[0030] When the target image is the first image, and a target number of the target image is same as a first number of the first image, it is determined that the verification is passed.

[0031] To achieve the above object, a second aspect of the embodiments of the present application provides an access verification method, the method comprising:

[0032] In response to an access instruction, sending a randomly generated prompt text to a server, so that the server feeds back at least one first image and at least one second image according to the prompt text, wherein the first image conforms to the description of the prompt text, and the second image does not conform to the description of the prompt text;

[0033] Receiving the first image and the second image sent by the server, and displaying a verification control, wherein a first area in the verification control displays the prompt text, and a second area in the verification control displays each of the first image and each of the second image;

[0034] In response to an image selection operation on the second area, determining a target image selected by the image selection operation;

[0035] Sending the target image to the server, so that the server determines a verification result according to the target image and the first image.

[0036] To achieve the above object, a third aspect of the embodiments of the present application provides an access verification device, comprising:

[0037] A generating unit, configured to receive a prompt text randomly generated and sent by an access device, and generate at least one first image and at least one second image according to the prompt text, wherein the first image conforms to the description of the prompt text, and the second image does not conform to the description of the prompt text;

[0038] A first sending unit, configured to send the first image and the second image to the access device, so that the prompt text is displayed in a first area in a verification control of the access device, and each of the first image and each of the second image is displayed in a second area in the verification control;

[0039] A receiving unit, configured to receive a target image sent by the access device, wherein the target image is an image determined by the access device according to an image selection operation on the second area;

[0040] A first determining unit, configured to determine a verification result based on the first image and the target image.

[0041] In some embodiments, the generating unit comprises:

[0042] The first determining sub-unit is configured to determine a first text vector corresponding to the prompt text.

[0043] The changing sub-unit is configured to change a target word in the prompt text that meets a changing condition to obtain a modified text.

[0044] The second determining sub-unit is configured to determine a second text vector corresponding to the modified text.

[0045] The first generating sub-unit is configured to generate at least one first image meeting the description of the prompt text based on the first text vector.

[0046] The second generating sub-unit is configured to generate at least one second image not meeting the description of the prompt text based on the second text vector.

[0047] In some embodiments, the changing sub-unit is configured to:

[0048] obtain a plurality of candidate words constituting the prompt text;

[0049] determine a word type corresponding to each of the candidate words;

[0050] determine a candidate word of a preset word type from the plurality of candidate words as a target word;

[0051] change at least one target word in the prompt text to a word of the same word type to obtain a modified text.

[0052] In some embodiments, the first determining sub-unit is configured to:

[0053] perform word segmentation processing on the prompt text to obtain a plurality of first words;

[0054] encode each of the first words to obtain a corresponding first word vector;

[0055] combine the plurality of first word vectors into a first text vector;

[0056] The second determining sub-unit is configured to:

[0057] perform word segmentation processing on the modified text to obtain a plurality of second words;

[0058] encode each of the second words to obtain a corresponding second word vector;

[0059] combine the plurality of second word vectors into a second text vector.

[0060] In some embodiments, the first determining unit is configured to:

[0061] When the target image is the first image, and the target number of the target image is the same as the first number of the first image, it is determined that the verification is passed.

[0062] To achieve the above object, a fourth aspect of the embodiments of the present application provides an access verification device, comprising:

[0063] A second sending unit is configured to send, in response to an access instruction, a randomly generated prompt text to a server, so that the server feeds back at least one first image and at least one second image according to the prompt text, wherein the at least one first image conforms to the description of the prompt text, and the at least one second image does not conform to the description of the prompt text.

[0064] A display unit is configured to receive the first image and the second image sent by the server, display a verification control, and display the prompt text in a first area in the verification control, and display each of the first image and the second image in a second area in the verification control.

[0065] A second determining unit is configured to determine, in response to an image selection operation on the second area, a target image selected by the image selection operation.

[0066] A third sending unit is configured to send the target image to the server, so that the server determines a verification result according to the target image and the first image.

[0067] To achieve the above object, a fifth aspect of the embodiments of the present application provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the method of the first aspect when executing the computer program.

[0068] To achieve the above object, a sixth aspect of the embodiments of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method of the first aspect.

[0069] The access verification method, device, electronic equipment and storage medium provided by the present application receive a prompt text randomly generated and sent by an access device, generate at least one first image meeting the description of the prompt text and at least one second image not meeting the description of the prompt text according to the prompt text, send the first image and the second image to the access device, so that the prompt text is displayed in a first area in a verification control of the access device, and each first image and each second image is displayed in a second area in the verification control, receive a target image sent by the access device, the target image being an image determined by the access device according to an image selection operation for the second area, and determine a verification result based on the first image and the target image. In this way, the first image and the second image that have never appeared before are generated according to the randomly generated prompt text in the verification process, so that the access verification is performed through the first image and the second image, the model for cracking the verification mechanism cannot learn the randomly generated image, the image verification code is prevented from being cracked due to image library leakage, and the verification security is improved. BRIEF DESCRIPTION OF DRAWINGS

[0070] Figure 1 is a flowchart of an access verification method provided by an embodiment of the present application;

[0071] Figure 2 is a display schematic diagram of a verification control provided by an embodiment of the present application;

[0072] Figure 3 is an operation schematic diagram of an image selection operation for a second area provided by an embodiment of the present application;

[0073] Figure 4 is a flowchart of step S101 in Figure 1

[0074] Figure 5 is a flowchart of step S202 in Figure 4

[0075] Figure 6 is a flowchart of step S201 in Figure 4

[0076] Figure 7 is a flowchart of step S104 in Figure 1

[0077] Figure 8 is another flowchart of an access verification method provided by an embodiment of the present application;

[0078] Figure 9 is a structural schematic diagram of an access verification device provided by an embodiment of the present application;

[0079] ​​​​ Figure 10 is another structural schematic diagram of an access verification device provided by an embodiment of the present application;

[0080] Figure 11 is a hardware structure schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0081] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0082] It should be noted that although the functional modules are divided in the device schematic diagram, and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a manner different from the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification and claims and the above-described drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence.

[0083] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0084] First, the meanings of several terms involved in the present application are analyzed:

[0085] Artificial Intelligence Generated Content (AIGC): mainly relying on artificial intelligence technology, especially deep learning algorithms. Neural networks in deep learning, such as generative adversarial networks (GAN), variational autoencoder (VAE), diffusion models, etc., provide strong technical support for AIGC. These models can learn patterns and features in large amounts of data, enabling them to generate new content similar to training data but not completely identical. It can be used for image generation, generating images based on text descriptions or specific requirements. For example, input "a red butterfly flying in the flower cluster", AIGC model can generate a vivid image showing the scene of red butterfly in the flower cluster. In addition, style transfer can be performed to apply the style of one image to another image, creating unique artistic effects.

[0086] AIGC can quickly generate a large amount of content, greatly improving the efficiency of creation. For some repetitive creation tasks, such as templated writing of news reports, batch generation of images, etc., AIGC can save a lot of time and labor costs. Provides new sources of inspiration and creativity for creators. The content generated by AIGC often has a unique style and perspective, which can stimulate the imagination of creators and encourage them to explore new directions of creation. According to the needs and preferences of users, it can be customized, such as generating exclusive music playlists, image wallpapers or story content for users, meeting the individual needs of different users. AIGC has broad application prospects in many industries, such as media, advertising, entertainment, education, etc. It can bring new business models and development opportunities to these industries, promoting the innovation and upgrading of the industry.

[0087] Based on this, the embodiment of the application provides an access verification method and device, electronic equipment and storage medium, aiming to avoid the model cracking of the image verification code cracking verification mechanism, and improve the verification security.

[0088] The access verification method, device, electronic equipment and storage medium provided by the embodiment of the application are specifically explained by the following embodiments. First, the access verification method in the embodiment of the application is described.

[0089] The access verification method provided by the embodiment of the application relates to the technical field of business processing. The access verification method provided by the embodiment of the application can be applied in a terminal, can be applied in a server end, and can also be software running in a terminal or a server end. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc.; the server end can be configured as an independent physical server, can be configured as a server cluster or a distributed system composed of multiple physical servers, can also be configured as a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and basic cloud computing services such as big data and artificial intelligence platforms; and the software can be an application that implements the access verification method, but is not limited to the above forms.

[0090] The application is operable in a variety of generic or specific computer system environments or configurations. For example: personal computer, server computer, handheld or portable device, tablet device, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like. The application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like, that perform particular tasks or implement particular abstract data types. The application can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are connected through a communication network. In a distributed computing environment, program modules can be located in both local and remote computer storage media including memory storage devices.

[0091] It should be noted that in various specific embodiments of the present application, when it is necessary to process relevant data related to the identity or characteristics of the user according to user information, user behavior data, user history data, and user location information, etc., the user's permission or consent will be obtained first, and the collection, use and processing of these data will comply with relevant laws, regulations and standards. In addition, when the embodiments of the present application need to obtain sensitive personal information of the user, the separate permission or separate consent of the user will be obtained through a pop-up window or a jump to a confirmation page, and after obtaining the separate permission or separate consent of the user, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0092] Figure 1 is an optional flowchart of the access verification method provided by the embodiments of the present application, Figure 1 The method in the above method can be applied to a server, and can include but not limited to steps S101 to S105.

[0093] Step S101, receiving a prompt text randomly generated and sent by an access device, generating at least one first image consistent with the description of the prompt text and at least one second image not consistent with the description of the prompt text according to the prompt text.

[0094] Wherein, the prompt text is a text randomly generated by the access device operated by the user when the user requests to access a certain website or log in to a personal webpage, which is used to describe a scene picture. The access device will send the randomly generated prompt text to the server, so that the server generates at least one first image consistent with the description of the scene picture in the prompt text and at least one second image not consistent with the description of the scene picture in the prompt text.

[0095] For example, the prompt text is "a Cocker Spaniel beside a window", after the server receives the prompt text, it generates two images of a Cocker Spaniel beside a window, one image of a Cocker Spaniel beside a door, and one image of a Ragdoll cat beside a window. Since the two images of a Cocker Spaniel beside a window are consistent with the description of the prompt text "a Cocker Spaniel beside a window", they are the first images. Since the image of a Cocker Spaniel beside a door and the image of a Ragdoll cat beside a window are not consistent with the description of the prompt text "a Cocker Spaniel beside a window", they are the second images.

[0096] In step S102, the first images and the second images are sent to the access device, so that the prompt text is displayed in the first area of the verification control of the access device, and each of the first images and each of the second images are displayed in the second area of the verification control.

[0097] In step S102, the first images and the second images are sent to the access device, so that the prompt text is displayed in the first area of the verification control of the access device, and each of the first images and each of the second images are displayed in the second area of the verification control.

[0098] Specifically, please refer to Figure 2 , Figure 2 The display schematic diagram of the verification control provided by the embodiment of the present application. Figure 2 The verification control 10 displayed in the display screen of the access device, and the verification control 10 includes a first area 11 and a second area 12. The prompt text is displayed in the first area 11, and each of the first images and each of the second images received by the access device from the server are displayed in the second area 12. The user selects the image consistent with the description of the prompt text from the multiple first images and the multiple second images displayed in the second area 12 according to the description in the prompt text.

[0099] The framework of the verification control will use HTML, CSS, JavaScript and related front-end frameworks (such as Vue.js) and other front-end technologies, so that it can adapt to and correctly display on various devices and screen sizes.

[0100] In step S103, the target image sent by the access device is received, and the target image is determined by the access device according to the image selection operation of the second area.

[0101] When the access device determines the target image selected by the user according to the image selection operation of the user for the second area, the target image is sent to the server so that the server verifies whether the target image is consistent with the generated first image according to the target image.

[0102] Specifically, refer to Figure 3 , Figure 3 The operation schematic diagram of the image selection operation for the second area provided by the embodiment of the present application. Taking the access device with a touch screen as an example, the image selection operation can be a click operation for the image in the second area. The user clicks the area where the image is located on the touch screen by a finger, so that the access device determines the display image corresponding to the clicked area and determines the display image as the target image. Taking the access device without a touch screen as an example, the image selection operation can be to determine the display image selected by an external device (keyboard, mouse, etc.) through the external device, and determine the display image as the target image.

[0103] In step S104, a verification result is determined based on the first image and the target image.

[0104] In step S103, after the server receives the target image sent by the access device, the target image is matched with the previously generated first image, so that it is determined whether the user access this time is verified, and the verification result is obtained.

[0105] Specifically, if the target image matches the previously generated first image, it is determined that the verification result of the user access this time is verified, and the user is allowed to access; if the target image does not match the previously generated first image, it is determined that the verification result of the user access this time is not verified, and the user is not allowed to access.

[0106] The steps S101 to S105 shown in the embodiments of the present application include the following steps: receiving a prompt text randomly generated and sent by an access device, generating at least one first image consistent with the description of the prompt text and at least one second image inconsistent with the description of the prompt text according to the prompt text; sending the first image and the second image to the access device, so that the prompt text is displayed in a first area in a verification control of the access device, and each of the first image and the second image is displayed in a second area in the verification control; receiving a target image sent by the access device, the target image being an image determined by the access device according to an image selection operation on the second area; and determining a verification result based on the first image and the target image. In this way, in the verification process, the first image and the second image that have never appeared before are generated according to the randomly generated prompt text, so that the access verification is performed through the first image and the second image, the model that cracks the verification mechanism cannot learn the randomly generated image, the image verification code is prevented from being cracked due to the leakage of the image library, and the verification security is improved.

[0107] Please refer to Figure 4 In some embodiments, step S101 can include but is not limited to steps S201 to S205:

[0108] Step S201: determining a first text vector corresponding to the prompt text.

[0109] In this way, when the server generates the first image and the second image according to the prompt text, the AIGC model is used for generation. For the generation of the first image, the AIGC model encodes the prompt text to obtain a first text vector corresponding to the prompt text, so as to generate the first image consistent with the description of the prompt text according to the first text vector subsequently.

[0110] Step S202: modifying a target word in the prompt text that meets a modification condition to obtain a modified text.

[0111] In this way, since the image generated by the trained AIGC model is highly consistent with the description of the text, in order to generate the second image inconsistent with the description of the prompt text, the original prompt text needs to be modified by word segmentation to form the modified text inconsistent with the original prompt text.

[0112] Specifically, since the modification of conjunctions such as “of” and “of” does not affect the description, the target word that meets the modification condition needs to be selected from the prompt text to modify the target word to obtain the modified text. The modification condition is that the description after the word segmentation modification is inconsistent with the description of the original prompt text.

[0113] Step S203, determine the second text vector corresponding to the modified text.

[0114] Wherein, after obtaining the modified text in step S202, the AIGC model encodes the modified text to obtain the second text vector corresponding to the modified text, so as to generate a second image inconsistent with the description of the prompt text according to the second text vector subsequently.

[0115] Step S204, generate at least one first image consistent with the description of the prompt text based on the first text vector.

[0116] Wherein, for the generation of the first image, the AIGC model will perform semantic understanding on the encoded first text vector. It extracts key feature information in the text based on the language knowledge and semantic relationship learned during pre-training. The AIGC model usually maps text features to a latent space. This latent space is a mathematical space predefined by the model, where each point represents a possible combination of image features. In this space, the text features related to "a corgi by the window" will be mapped to a specific area, and the vectors in this area represent the feature combinations that can generate images consistent with the description. For example, in this latent space, the feature vector of the corgi and the feature vector of the window scene will be combined in a specific way.

[0117] Based on the mapping results in the latent space, the AIGC model begins to generate images. It uses techniques such as generative adversarial networks (GAN) or diffusion models (Diffusion Model) for image generation.

[0118] In the diffusion model, the model starts from a random noise image, gradually removes noise and adds image features corresponding to the text description. For example, first determine the general outline of the corgi, then gradually refine its body features, fur color, and add window scene elements, etc., to finally generate an image consistent with the text description.

[0119] Step S205, generate at least one second image that does not conform to the description of the prompt text based on the second text vector.

[0120] For the generation of the second image, the AIGC model will perform semantic understanding on the encoded second text vector. It extracts key feature information in the text based on the language knowledge and semantic relationships learned during pre-training. The AIGC model usually maps the text features to a latent space. This latent space is a mathematical space predefined by the model, where each point represents a possible combination of image features. In this space, the text features related to "a corgi by the window" will be mapped to a specific region, and the vectors in this region represent the feature combinations that can generate images consistent with the description. For example, in this latent space, the feature vector of a corgi and the feature vector of a window scene will be combined in a specific way.

[0121] Based on the mapping results in the latent space, the AIGC model begins to generate images. It uses techniques such as Generative Adversarial Networks (GAN) or Diffusion Models to generate images.

[0122] In GAN, the generator network generates images based on the feature vectors obtained from the latent space, and the discriminator network then discriminates between the generated images. The generator continuously adjusts its parameters to generate more realistic and textually consistent images, and the discriminator continuously improves its discrimination ability. In this adversarial process, the quality of the generated images is continuously improved until the required image (such as a corgi by the window) is generated.

[0123] In this way, by generating the first image consistent with the prompt text description based on the first text vector and the second image inconsistent with the prompt text description based on the second text vector, the complexity and diversity of the verification code are increased. This makes it difficult for malicious programs to predict and crack the verification code, improving the security of the system. Moreover, because traditional verification codes are easily recognized by machines, the images generated by this method have higher randomness and complexity, especially when combined with different text vectors and advanced generation techniques such as GAN and Diffusion Models, making automated attacks more difficult.

[0124] Please refer to Figure 5 In some embodiments, step S202 can include but is not limited to steps S301 to S304:

[0125] Step S301, obtaining a plurality of candidate words constituting the prompt text.

[0126] When given a prompt text, it needs to use word segmentation technology in natural language processing to split it into independent words, which are candidate words.

[0127] For example, the prompt text is "a lovely cat in the garden", first the text is segmented. The sentence can be split into individual words using a segmentation tool in natural language processing, resulting in multiple candidate segments, i.e. "one", "lovely", "of", "cat", "in", "garden", "in".

[0128] Step S302, determine the type of each candidate segment.

[0129] Different words have different roles in grammar and semantics, and can be classified into different types. Common segmentation types include nouns, verbs, adjectives, adverbs, prepositions, and auxiliary words.

[0130] For example, the prompt text is "a lovely cat in the garden", "one" can be determined as a quantifier type; "lovely" is an adjective type; "of" is an auxiliary word type; "cat" is a noun type; "in" is a preposition type; "garden" is a noun type; "in" is a direction word type.

[0131] Step S303, determine the target segment from the candidate segment whose segmentation type is the preset segmentation type.

[0132] Wherein, the preset segmentation type is a type of segmentation that can form a text inconsistent with the original text after being changed.

[0133] For example, set the preset segmentation type as noun and adjective. Then in the prompt text "a lovely cat in the garden", "lovely" (adjective type) and "cat" (noun type), "garden" (noun type) will be determined as target segment.

[0134] Step S304, change at least one of the target segments in the prompt text to a segment of the same segmentation type to obtain a modified text.

[0135] Wherein, after determining the target segment in step S303, at least one of the target segments in the prompt text is changed to a segment of the same segmentation type to obtain a modified text.

[0136] For example, the target segment in the prompt text "a lovely cat in the garden" is "cat", then the target segment "cat" is changed. Since "cat" is a noun type, we can choose a word from the same segmentation type (noun) to replace it, such as replacing it with "dog", then the modified text becomes "a lovely dog in the garden". In this way, a modified text inconsistent with the original prompt text is obtained.

[0137] In this way, by performing word segmentation processing on the prompt text and determining the target word segmentation for modification, a large number of different modified texts can be generated, thereby increasing the diversity and randomness of the verification code image. It is difficult for malicious programs to predict and crack this dynamically changing verification code, thereby improving the security of the system. Compared with traditional verification codes that are easily recognized by automated programs, this method based on natural language processing and flexible word segmentation modification makes the verification code more complex. Different word segmentation types and modifiable target word segmentation make it difficult for automated attacks to find a fixed pattern for cracking.

[0138] Referring to Figure 6 In some embodiments, step S201 can include but is not limited to steps S401-S403, and step S203 can include but is not limited to steps S404-S406:

[0139] Step S401, performing word segmentation processing on the prompt text to obtain a plurality of first word segments.

[0140] Since the prompt text is composed of a plurality of words, the word segmentation processing on the prompt text obtains a plurality of first word segments that constitute the prompt text.

[0141] Specifically, the word segmentation processing can be performed in a dictionary-based manner, specifically: starting from the left of the text to be segmented, a string of a certain length (e.g., seven characters) is taken as a matching field, and it is checked whether this word exists in the dictionary. If it exists, the field is cut out as a word; if it does not exist, the last character of the matching field is removed and the search continues in the dictionary until a word is found or only one character remains.

[0142] For example, for the text "I like to eat apples", suppose the dictionary has the words "I", "like", "eat", "apple", etc. First, take "I like to eat apples" as a matching field, which is not found in the dictionary, then remove the last character to become "I like to eat", which is still not found, and then remove another character to become "I like to eat", and so on, until the word "I" is found and cut out. Then continue processing the remaining text "like to eat apples", and finally get the word segments "I", "like", "eat", and "apple".

[0143] Step S402, encoding each first word segment to obtain a corresponding first word segment vector.

[0144] A pre-trained word embedding model (such as Word2Vec or BERT and related variants) can be used to map each word to a vector in a high-dimensional vector space.

[0145] For example, for the text "a Corgi by the window", the words "Corgi", "by the window", etc. will be converted into corresponding vectors, which contain the semantic information of the words.

[0146] Step S403, combining the plurality of first word segmentation vectors into a first text vector.

[0147] After determining the first word segmentation vector corresponding to each first word segmentation in step S402, the word segmentation vector of each word segmentation can be sequentially spliced together in the order of appearance in the text to form a long vector as a text vector.

[0148] For example, for the text "a cute cat in the garden", after word vector representation, the vector of "a" is [0.1], the vector of "cute" is [0.4], the vector of "of" is [0.7], the vector of "cat" is [1.0], the vector of "in" is [1.3], the vector of "garden" is [1.6], and the vector of "in" is [1.9]. The text vector obtained by sequentially splicing these vectors is [0.1, 0.4, 0.7, 1.0, 1.3, 1.6, 1.9].

[0149] Alternatively, the first text vector can be obtained by weighted summation of each first word segmentation, specifically: the weight of each word segmentation in the text vector is determined according to the importance of each word segmentation. For example, Term Frequency-Inverse Document Frequency (TF-IDF) can be used to measure the importance of each word segmentation. The higher the TF-IDF value of a word segmentation, the higher the weight of the word segmentation in the text vector can be set. In addition, the attention mechanism in deep learning can also be used to automatically learn the weight of each word segmentation. The attention mechanism can dynamically adjust the importance of each word segmentation according to the current task and context information. For the word segmentation vector of each word segmentation, multiply the corresponding weight and sum up to obtain the text vector.

[0150] For example, assuming the vector of "a" is [0.1, 0.2, 0.3] and the weight is 0.2; the vector of "cute" is [0.4, 0.5, 0.6] and the weight is 0.3; the vector of "of" is [0.7, 0.8, 0.9] and the weight is 0.1; the vector of "cat" is [1.0, 1.1, 1.2] and the weight is 0.4; the vector of "in" is [1.3, 1.4, 1.5] and the weight is 0.1; the vector of "garden" is [1.6, 1.7, 1.8] and the weight is 0.3; the vector of "in" is [1.9, 2.0, 2.1] and the weight is 0.1.

[0151] The text vector = 0.2 * [0.1, 0.2, 0.3] + 0.3 * [0.4, 0.5, 0.6] + 0.1 * [0.7, 0.8, 0.9] + 0.4 * [1.0, 1.1, 1.2] + 0.1 * [1.3, 1.4, 1.5] + 0.3 * [1.6, 1.7, 1.8] + 0.1 * [1.9, 2.0, 2.1].

[0152] The calculated text vector is a three-dimensional vector, and its specific value depends on the values of each word vector and weight. This method considers the importance of each word, can more effectively combine word vectors, and can adapt to different tasks and datasets through different weight determination methods.

[0153] Step S404, performing word segmentation processing on the modified text to obtain a plurality of second words.

[0154] Among them, for the modified text, the modified text is also subjected to word segmentation processing to obtain a plurality of second words constituting the modified text.

[0155] Specifically, the word segmentation processing method can also be: using an N-gram language model for word segmentation, N-gram is a word segmentation method based on statistical language model, which considers the appearance probability of continuous N characters or words in the text. Usually, N can be 2 (bi-gram), 3 (tri-gram), etc.

[0156] For example, for the text "today the weather is good", if a bi-gram model is used, the appearance probability of adjacent two words such as "today", "weather", "very good" will be counted. When word segmentation, according to these probabilities to determine the best word segmentation position. If the probability of "today" and "weather" appearing together is high, and the probability of "weather" and "very good" appearing together is also high, then the sentence can be divided into "today", "weather", "very good" three words.

[0157] Step S405, encoding each second word to obtain a corresponding second word vector.

[0158] Wherein, the way of encoding each second word is the same as the way of encoding each first word, which is not described here.

[0159] Step S406, combining a plurality of second word vectors into a second text vector.

[0160] Wherein, the way of combining a plurality of second word vectors into a second text vector is the same as the way of combining a plurality of first word vectors into a first text vector, which is not described here.

[0161] In this way, the first text vector and the second text vector obtained through the above steps can accurately reflect the semantic information of the prompt text and the modified text. When these text vectors are passed as input to the image generation model (such as an AIGC model), they can provide more explicit guidance for image generation, making the generated images more consistent with the text description. For example, for the prompt text "a Corgi dog by the window", the processed text vector can guide the image generation model to generate an image with Corgi dog features and a window scene; and for the modified text, such as "a Husky by the door", the corresponding text vector can prompt the model to generate different image content.

[0162] Please refer to Figure 7 In some embodiments, step S104 can include but is not limited to including step S501:

[0163] Step S501, when the target image is the first image, and the target number of the target image is the same as the first number of the first image, it is determined that the verification is passed.

[0164] Wherein, when the server receives the target image sent by the access device, it will match with the first image generated in advance. If the target image is the first image, and the target number of the target image is the same as the first number of the first image, it means that the user has selected all the first images according to the description of the prompt text, and then the verification is passed.

[0165] In this way, by requiring the target image to be completely consistent with the first image and the number to be the same, it is ensured that the user's selection is accurate and meets the description of the prompt text. This strict matching method can effectively prevent users from misselecting or randomly selecting images, improving the accuracy of verification. For example, when the prompt text is "three red apples in a basket", only when the user accurately selects all three apple images that meet the description can it be considered as verification passed, avoiding false judgments caused by ambiguous selection or incorrect selection.

[0166] Figure 8 is an optional flowchart of the access verification method provided by the embodiments of the present application, Figure 8 The method in the above method applied to the access device can include but is not limited to including steps S601 to S604:

[0167] Step S601, in response to an access instruction, a randomly generated prompt text is sent to a server to make the server feedback at least one first image consistent with the description of the prompt text and at least one second image not consistent with the description of the prompt text according to the prompt text.

[0168] The access instruction is a login operation of a user when accessing a website. After the access device receives the login operation, it is equivalent to responding to the access instruction of the user. A randomly generated prompt text is generated and sent to the server. The server feeds back at least one first image and at least one second image that do not meet the description of the prompt text according to the prompt text.

[0169] In step S602, the first image and the second image sent by the server are received, and a verification control is displayed. The first area in the verification control displays the prompt text, and the second area in the verification control displays each of the first image and each of the second image.

[0170] The access device displays the prompt text in the first area of the verification control of the display screen, and displays each of the first image and each of the second image received from the server in the second area.

[0171] In step S603, in response to an image selection operation on the second area, a target image selected by the image selection operation is determined.

[0172] The user selects a target image consistent with the description of the prompt text from the multiple images displayed in the second area according to the prompt text displayed in the first area of the verification control, so that the verification control determines the target image selected by the user through the image selection operation.

[0173] In step S604, the target image is sent to the server, so that the server determines a verification result according to the target image and the first image.

[0174] The access device sends the determined target image to the server to trigger the server to determine a verification result according to the target image and the first image.

[0175] In this way, by randomly generating the prompt text, the image content in each verification process is different, which increases the difficulty of prediction and cracking of malicious programs. It is difficult for an automated attack program to prepare a cracking strategy in advance for a specific prompt text, thereby improving the security of the system. For example, hackers cannot predict what kind of prompt text and image combination will appear, greatly reducing the risk of being attacked.

[0176] Please refer to Figure 9 The embodiments of the present application also provide an access verification device applied to a server and capable of implementing the above access verification method. The device comprises:

[0177] The generating unit is configured to receive prompt text randomly generated and sent by the access device, generate at least one first image consistent with the prompt text and at least one second image inconsistent with the prompt text according to the prompt text;

[0178] The first sending unit is configured to send the first image and the second image to the access device, so that the prompt text is displayed in a first region in a verification control of the access device, and each of the first image and each of the second image is displayed in a second region in the verification control;

[0179] The receiving unit is configured to receive a target image sent by the access device, the target image being an image determined by the access device according to an image selection operation on the second region;

[0180] The first determining unit is configured to determine a verification result based on the first image and the target image.

[0181] In some embodiments, the generating unit comprises:

[0182] The first determining sub-unit is configured to determine a first text vector corresponding to the prompt text;

[0183] The changing sub-unit is configured to change a target word in the prompt text that meets a changing condition to obtain a modified text;

[0184] The second determining sub-unit is configured to determine a second text vector corresponding to the modified text;

[0185] The first generating sub-unit is configured to generate at least one first image consistent with the prompt text based on the first text vector;

[0186] The second generating sub-unit is configured to generate at least one second image inconsistent with the prompt text based on the second text vector.

[0187] In some embodiments, the changing sub-unit is configured to:

[0188] Obtain a plurality of candidate words constituting the prompt text;

[0189] Determine a word type corresponding to each of the candidate words;

[0190] Determine a candidate word with a preset word type from the plurality of candidate words as a target word;

[0191] Change at least one target word in the prompt text to a word with the same word type to obtain a modified text.

[0192] In some embodiments, the first determining sub-unit is configured to:

[0193] perform word segmentation on the prompt text to obtain a plurality of first segmented words;

[0194] encode each of the first segmented words to obtain a corresponding first segmented word vector;

[0195] combine the plurality of first segmented word vectors into a first text vector;

[0196] The second determining unit is configured to:

[0197] perform word segmentation on the modified text to obtain a plurality of second segmented words;

[0198] encode each of the second segmented words to obtain a corresponding second segmented word vector;

[0199] combine the plurality of second segmented word vectors into a second text vector.

[0200] In some embodiments, the first determining unit is configured to:

[0201] When the target image is the first image and the target number of the target image is the same as the first number of the first image, it is determined that the verification is passed.

[0202] Referring to Figure 10 The embodiments of the present application also provide an access verification device applied to an access equipment, which can implement the above access verification method. The device comprises:

[0203] The second sending unit is configured to, in response to an access instruction, send a randomly generated prompt text to a server, so that the server feeds back at least one first image meeting the description of the prompt text and at least one second image not meeting the description of the prompt text according to the prompt text;

[0204] The display unit is configured to receive the first image and the second image sent by the server, display a verification control, and display the prompt text in a first area in the verification control and display each of the first image and the second image in a second area in the verification control.

[0205] The second determining unit is configured to, in response to an image selection operation on the second area, determine a target image selected by the image selection operation.

[0206] The third sending unit is configured to send the target image to the server, so that the server determines a verification result according to the target image and the first image.

[0207] The access verification device provided by the embodiments of the present application receives prompt text randomly generated and sent by an access device through a generation unit, generates at least one first image conforming to the description of the prompt text and at least one second image not conforming to the description of the prompt text according to the prompt text; a first sending unit sends the first image and the second image to the access device, so that the prompt text is displayed in a first area in a verification control of the access device, and each first image and each second image is displayed in a second area in the verification control; a receiving unit receives a target image sent by the access device, the target image being an image determined by the access device according to an image selection operation on the second area; and a first determination unit determines a verification result based on the first image and the target image. In this way, in the verification process, the first image and the second image that have never appeared before are generated according to the randomly generated prompt text, so that the access verification is performed through the first image and the second image, the model that cracks the verification mechanism cannot learn the randomly generated image, the image verification code is prevented from being cracked due to image library leakage, and the verification security is improved. By generating the first image conforming to the description of the prompt text based on the first text vector and generating the second image not conforming to the description of the prompt text based on the second text vector, the complexity and diversity of the verification code are increased. This makes it difficult for malicious programs to predict and crack the verification code, improving the security of the system. Moreover, the images generated by the traditional verification code are easily recognized by machines, while the images generated by the method have higher randomness and complexity, especially in combination with different text vectors and advanced generation techniques (such as generative adversarial networks and diffusion models), making it more difficult for automated attacks. By performing word segmentation processing on the prompt text and determining the target word segmentation for modification, a large number of different modified texts can be generated, thereby increasing the diversity and randomness of the verification code images. It is difficult for malicious programs to predict and crack this dynamically changing verification code, improving the security of the system. Compared with the traditional verification code that is easily recognized by automated programs, this method based on natural language processing and flexible word segmentation modification makes the verification code more complex. Different word segmentation types and changeable target word segmentation make it difficult for automated attacks to find a fixed pattern for cracking. The first text vector and the second text vector obtained through the above steps can accurately reflect the semantic information of the prompt text and the modified text. When these text vectors are input into the image generation model (such as the AIGC model), they can provide more explicit guidance for image generation, making the generated images more consistent with the text description. For example, for the prompt text "a Corgi dog by the window", the text vector obtained after processing can guide the image generation model to generate an image with Corgi dog features and a window scene; and for the modified text, such as "a Siberian husky by the door", the corresponding text vector can prompt the model to generate different image content.By requiring the target images to be identical to the first image and the same number, it is ensured that the user's selection is accurate and meets the description of the prompt text. This strict matching method can effectively prevent users from misselecting or randomly selecting images, improving the accuracy of verification. For example, when the prompt text is "three red apples in a basket", only when the user accurately selects all three apple images that meet the description will the verification be considered successful, avoiding false positives due to ambiguous selection or incorrect selection. By randomly generating prompt texts, the image content in each verification process is different, increasing the difficulty of malicious program prediction and cracking. Automated attack programs are difficult to prepare cracking strategies in advance for specific prompt texts, thereby improving the security of the system. For example, hackers cannot predict what prompt texts and image combinations will appear, greatly reducing the risk of being attacked.

[0208] The specific implementation of the access verification apparatus is basically the same as the specific embodiments of the above access verification method, and will not be repeated here.

[0209] The embodiments of the present application also provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor implements the above access verification method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0210] Please refer to Figure 11 , Figure 11 The hardware structure of the electronic device of another embodiment is illustrated, which includes:

[0211] The processor 901 can be implemented in the form of a general-purpose CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, etc., for executing related programs to implement the technical solutions provided by the embodiments of the present application;

[0212] The memory 902 can be implemented in the form of a ROM (ReadOnly Memory), a static storage device, a dynamic storage device, or a RAM (Random Access Memory), etc. The memory 902 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present application are implemented by software or firmware, the related program codes are saved in the memory 902 and called and executed by the processor 901 to implement the access verification method of the embodiments of the present application;

[0213] The input / output interface 903 is used to realize information input and output;

[0214] The communication interface 904 is configured to realize the communication interaction between the device and other devices, and can realize the communication through a wired manner (for example, a USB, a network cable, or the like) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, or the like).

[0215] The bus 905 is configured to transmit information between various components (for example, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904) of the device.

[0216] The processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are connected to each other through the bus 905 to realize the communication connection between the device.

[0217] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the access verification method.

[0218] The memory is a non-transitory computer readable storage medium, and can be used to store a non-transitory software program and a non-transitory computer executable program. In addition, the memory can include a high-speed random access memory, and can further include a non-transitory memory, for example, at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor, and the remote memory can be connected to the processor through a network. Examples of the network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0219] The access verification method, the access verification device, the electronic device, and the storage medium provided by the embodiment of the present application can receive a prompt text randomly generated and sent by an access device, generate at least one first image conforming to the prompt text description and at least one second image not conforming to the prompt text description according to the prompt text, send the first image and the second image to the access device, so that the prompt text is displayed in a first region in a verification control of the access device, and each first image and each second image is displayed in a second region in the verification control, receive a target image sent by the access device, the target image is an image determined by the access device according to an image selection operation of the second region, and determine a verification result based on the first image and the target image. In this way, the first image and the second image that have never appeared are generated according to the randomly generated prompt text in the verification process, the access verification is performed through the first image and the second image, a model for cracking the verification mechanism cannot learn the randomly generated image, the image verification code is prevented from being cracked due to the leakage of the image library, and the verification security is improved.

[0220] The embodiments described in the specification are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0221] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than shown in the figures, or combine certain steps, or different steps.

[0222] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separate, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0223] Those skilled in the art can understand that all or some of the steps in the above disclosed method, the functional modules / units in the system and the device can be implemented as software, firmware, hardware and their appropriate combinations.

[0224] The terms "first", "second", "third", "fourth" and the like (if any) in the specification of the present application and the above-described figures are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0225] It should be understood that, in the application, "at least one" refers to one or more, and "multiple" refers to two or more. "And / or" is used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, "A and / or B" can represent three cases of only A, only B, and A and B existing at the same time, wherein A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after it. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can represent a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, and c can be single or multiple.

[0226] In several embodiments provided in the application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the above units is only a logical function division, and actual implementation can have another division manner, 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 displayed or discussed mutual ones can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0227] The units described above 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 multiple network units. According to actual needs, some or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0228] In addition, each functional unit in each embodiment of the application can be integrated in one processing unit, or each unit can exist physically, 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 functional unit.

[0229] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in other words, the part that contributes to the prior art or the whole 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 multiple 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 of the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program storage media.

[0230] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. An access verification method, characterized in that: include: receiving a prompt text randomly generated and sent by an access device, and determining a first text vector corresponding to the prompt text; Obtaining multiple candidate word segments constituting the prompt text; Determine the segmentation type corresponding to each candidate segmentation; Determine a candidate participle whose participle type is a preset participle type as a target participle; Changing at least one target segmentation in the prompt text to a segmentation of the same segmentation type to obtain a modified text; Determining a second text vector corresponding to the modified text; generating at least one first image that matches the description of the prompt text based on the first text vector; generating at least one second image that does not conform to the description of the prompt text based on the second text vector; Sending the first image and the second image to the access device, so that the prompt text is displayed in a first area within the verification control of the access device, and each of the first image and each of the second images is displayed in a second area within the verification control; receiving a target image sent by the access device, where the target image is an image determined by the access device according to an image selection operation for the second area; A verification result is determined based on the first image and the target image.

2. The access verification method according to claim 1, characterized in that: The determining the first text vector corresponding to the prompt text includes: Performing word segmentation processing on the prompt text to obtain multiple first word segmentations; Encode each of the first participles to obtain a corresponding first participle vector; Combining the plurality of first word segmentation vectors into a first text vector; The determining the second text vector corresponding to the modified text includes: Performing word segmentation processing on the modified text to obtain a plurality of second word segmentations; Encode each of the second participles to obtain a corresponding second participle vector; Combine the plurality of the second word segmentation vectors into a second text vector.

3. The access verification method according to claim 1, wherein: The determining a verification result based on the first image and the target image includes: When the target image is the first image and the target number of the target image is the same as the first number of the first image, it is determined that the verification is passed.

4. An access verification method, characterized in that: include: In response to the access instruction, sending a randomly generated prompt text to a server, so that the server determines a first text vector corresponding to the prompt text; Obtaining multiple candidate word segments constituting the prompt text; Determine the segmentation type corresponding to each candidate segmentation; Determine a candidate participle whose participle type is a preset participle type as a target participle; Changing at least one target segmentation in the prompt text to a segmentation of the same segmentation type to obtain a modified text; Determining a second text vector corresponding to the modified text; generating at least one first image that matches the description of the prompt text based on the first text vector; generating at least one second image that does not conform to the description of the prompt text based on the second text vector, and feeding back at least one first image that conforms to the description of the prompt text and at least one second image that does not conform to the description of the prompt text; receiving the first image and the second image sent by the server, and displaying a verification control, wherein a first area within the verification control displays the prompt text, and a second area within the verification control displays each of the first image and each of the second images; In response to an image selection operation on the second area, determining a target image selected by the image selection operation; The target image is sent to the server, so that the server determines a verification result according to the target image and the first image.

5. An access verification device, characterized in that: include: The generating unit is configured to receive a prompt text randomly generated and sent by the access device, including: A first determining subunit, configured to determine a first text vector corresponding to the prompt text; Change the subunits for: Obtaining multiple candidate word segments constituting the prompt text; Determine the segmentation type corresponding to each candidate segmentation; Determine a candidate participle whose participle type is a preset participle type as a target participle; Changing at least one target segmentation in the prompt text to a segmentation of the same segmentation type to obtain a modified text; A second determining subunit, configured to determine a second text vector corresponding to the modified text; A first generating subunit, configured to generate at least one first image that conforms to the description of the prompt text based on the first text vector; a second generating subunit, configured to generate at least one second image that does not conform to the description of the prompt text based on the second text vector; a first sending unit, configured to send the first image and the second image to the access device, so that the prompt text is displayed in a first area within the verification control of the access device, and each of the first image and each of the second images is displayed in a second area within the verification control; a receiving unit, configured to receive a target image sent by the access device, wherein the target image is an image determined by the access device according to an image selection operation for the second area; The first determining unit is configured to determine a verification result based on the first image and the target image.

6. An access verification device, characterized in that: include: a second sending unit, configured to send a randomly generated prompt text to a server in response to the access instruction, so that the server determines a first text vector corresponding to the prompt text; Obtaining multiple candidate word segments constituting the prompt text; Determine the segmentation type corresponding to each candidate segmentation; Determine a candidate participle whose participle type is a preset participle type as a target participle; Changing at least one target segmentation in the prompt text to a segmentation of the same segmentation type to obtain a modified text; Determining a second text vector corresponding to the modified text; generating at least one first image that matches the description of the prompt text based on the first text vector; generating at least one second image that does not conform to the description of the prompt text based on the second text vector, and feeding back at least one first image that conforms to the description of the prompt text and at least one second image that does not conform to the description of the prompt text; a display unit, configured to receive the first image and the second image sent by the server, and display a verification control, wherein a first area within the verification control displays the prompt text, and a second area within the verification control displays each of the first image and each of the second images; a second determining unit, configured to determine, in response to the image selection operation on the second area, a target image selected by the image selection operation; A third sending unit is configured to send the target image to the server, so that the server determines a verification result according to the target image and the first image.

7. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the access verification method according to any one of claims 1 to 3 or the access verification method according to claim 4 when executing the computer program.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the access verification method according to any one of claims 1 to 3 or the access verification method according to claim 4 is implemented.

Images