AI-based one-click business card style changing method, device, and electronic device

Abstract

The embodiments of the present invention propose an AI-based method, device, and electronic device for changing the style of a business card with one click. The method receives a user-triggered one-click change of the business card style operation for the current business card, determines the current style of the current business card and the current user's information, matches multiple alternative styles in a predetermined style library based on the current user's information, uses the difference between the characteristic values ​​of the multiple alternative styles and the characteristic value of the current style as a style deviation, inputs the style deviation and the current business card into a generative large model, and obtains one or more recommended business cards; and displays the multiple recommended business cards on a style change recommendation interface. The above solution can significantly improve the personalization of business cards and has high generation efficiency.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an AI-based method, device, and electronic device for changing business card styles with one click. Background Art

[0002] Changing the business card style is usually done in the following way: obtaining the user's personal information and generating a QR code image based on the personal information; screening at least one candidate template from the preset template set for the user to select based on the age, gender and / or position in the personal information; obtaining a selection instruction, and determining the candidate template corresponding to the selection instruction as a personalized template; and generating a personalized business card based on the QR code image and the personalized template.

[0003] However, the personalization of this method is limited by the richness of templates. Generally speaking, the number of templates is limited, the probability of using the same template is relatively high, and the use of the same template may also result in similar situations. Summary of the Invention

[0004] In view of this, the object of the present invention is to provide a method, device and electronic device for changing the business card style with one click based on AI.

[0005] In order to achieve the above objectives, the technical solutions adopted in the embodiments of the present invention are as follows:

[0006] In a first aspect, the present invention provides an AI-based method for changing the style of a business card with one click, comprising: receiving a one-click business card style change operation triggered by a user for a current business card, determining the current style of the current business card and the information of the current user; matching multiple alternative styles in a predetermined style library based on the information of the current user; taking the difference between the feature values ​​of the multiple alternative styles and the feature value of the current style as a style deviation; inputting the style deviation and the current business card into a generative large model to obtain one or more recommended styles of business cards; and displaying multiple recommended styles of business cards on a style change recommendation interface.

[0007] In a second aspect, the present invention provides an AI-based device for changing the style of a business card with one click, comprising: a receiving module for receiving a one-click business card style change operation triggered by a user for a current business card, and determining the current style of the current business card and the information of the current user; a matching module for matching multiple alternative styles in a predetermined style library based on the information of the current user; a determination module for taking the difference between the feature values ​​of the multiple alternative styles and the feature value of the current style as a style deviation; a generation module for inputting the style deviation and the current business card into a generative large model to obtain one or more recommended styles of business cards; a recommendation module for displaying multiple recommended styles of business cards on a style change recommendation interface.

[0008] In a third aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement the method described in any one of the aforementioned embodiments.

[0009] The AI-based one-click business card style change method, device, and electronic device provided by the embodiments of the present invention determine the current style of the current business card and the current user's information by receiving a one-click business card style change operation triggered by the user for the current business card; match multiple alternative styles in a predetermined style library based on the current user's information; determine a style deviation based on the multiple alternative styles and the current style; input the style deviation and the current business card into a generative large model to obtain one or more recommended business cards; and display the multiple recommended business cards in a style change recommendation interface. The above scheme can significantly improve the personalization of business cards, and has high generation efficiency. The generated business card style not only includes the style in the template, but also generates richer styles with the help of the generative large model's capabilities, thereby reducing the probability of duplication of business card styles and improving differentiation.

[0010] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0012] Figure 1 A schematic structural diagram of an electronic device provided by an embodiment of the present invention is shown;

[0013] Figure 2 A flowchart of a method for changing the business card style with one click based on AI provided in an embodiment of the present application;

[0014] Figure 3 This is a functional module diagram of an AI-based device for changing business card styles with one click, provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0015] The following will be combined with the accompanying drawings to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0016] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but is merely intended to represent selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative work are within the scope of protection of the present invention.

[0017] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0018] Please refer to Figure 1 is a block diagram of an electronic device. The electronic device includes a memory, a processor, and a communication module. The memory, processor, and communication module are electrically connected to each other directly or indirectly to enable data transmission or exchange. For example, these components may be electrically connected to each other via one or more communication buses or signal lines.

[0019] Electronic device 100 includes a processor (CPU) 120, which can perform various appropriate actions and processes according to the programs stored in memory 110. Memory 110 may include read-only memory (ROM) or random access memory (RAM). Various programs and data required for the operation of electronic device 100 are also stored in RAM. The CPU, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0020] The following components are connected to the I / O interface: an input section including a keyboard, mouse, and the like; an output section including a cathode ray tube (CRT), liquid crystal display (LCD), and speakers; a storage section including a hard disk; and a communication module 130 including a network interface card such as a LAN card or modem. Communication module 130 performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. Removable media such as magnetic disks, optical disks, magneto-optical disks, and semiconductor memories are installed in the drive as needed, allowing computer programs read from these media to be installed in the storage section as needed.

[0021] In particular, according to embodiments of the present disclosure, the processes described below with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer-readable medium carrying instructions. In such embodiments, the instructions can be downloaded and installed from a network via a communication component and / or installed from a removable medium. When the instructions are executed by a CPU, the various method steps described herein are performed.

[0022] In some embodiments, the memory 110 is used to store programs or data. The memory may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0023] The processor 120 is used to read / write data or programs stored in the memory and execute corresponding functions.

[0024] The communication module 130 is used to establish a communication connection between the electronic device and other communication terminals through a network, and to send and receive data through the network.

[0025] It should be understood that Figure 1 The structure shown is only a schematic diagram of the structure of the electronic device, and the electronic device may also include Figure 1 More or fewer components than shown, or with Figure 1 Different configurations shown. Figure 1 Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0026] Figure 2 A flowchart of a method for changing the business card style with one click based on AI provided in an embodiment of the present application. The method can be based on Figure 1 The electronic device shown is implemented as Figure 2 As shown, the method may include:

[0027] S210: Receive a one-key change of business card style operation triggered by the user for the current business card, and determine the current style of the current business card and information of the current user.

[0028] The style of a business card can include fonts, color schemes, and layouts. The style of a business card refers to the visual presentation of the card, including font size and type, color schemes, and layout. These styles can directly affect the aesthetics and practicality of the card.

[0029] User information may include user personal information, user behavior information, and user scenario information, etc.

[0030] User personal information can refer to data directly related to the user, such as name, position, company name, etc. This information is usually used to display the core content of the business card, and can also be used to analyze the user's preference for style.

[0031] User behavior information can include browsing behavior, preferences, and other information. For example, it can include records of user activity within the system, such as browsing history and click preferences. This data can help the system better understand user needs and provide more tailored services.

[0032] The user's contextual information can refer to their current environment or status, such as their geographic location (whether they are in a meeting) or system interface information. Users may have multiple roles, performing different roles in different systems. The user's current role can be determined based on the system interface they are in. This information helps dynamically adjust the business card style to suit different contexts.

[0033] In some embodiments, when a user triggers a "one-click style change" action (e.g., clicking an interactive button or using a gesture), the OCR engine can extract the location and hierarchical relationships (title / body / contact information) of the business card text area and calculate the layout feature vector (e.g., information density, alignment). A CNN model (e.g., ResNet-50) is used to analyze the current business card's font type (via a text recognition model), color scheme (HSV color space histogram), and background texture (frequency domain analysis). In conjunction with an NLP model (e.g., BERT), entity recognition is performed on the business card content, annotating industry attributes (e.g., "finance / technology") and position level (using job keywords to match the corporate VI standard library).

[0034] You can also extract structured data such as name, position, company, industry, etc. from the user database.

[0035] You can also count the frequency of business card edits in the past 30 days (such as the number of color changes / layout switching rate) to build a behavioral time series.

[0036] The user's current scenario (such as "mobile office / business meeting") can also be inferred through device sensors (GPS positioning, gyroscope tilt angle) and network status (Wi-Fi SSID).

[0037] You can also call the enterprise API in real time to obtain the user's recent business dynamics (such as the industry of newly signed customers) as temporary scene feature injection.

[0038] In addition, the current style may also be pre-configured information. The current user's information may also be based on the user's account information or obtained from a human resources system.

[0039] For example, user account information is securely synchronized with the enterprise's unified authentication center via the OAuth 2.0 protocol, with incremental data updates performed hourly. The HR system integrates data using the SOAP interface of the Enterprise Service Bus (ESB). XPath parsing of XML-formatted employee profile data automatically desensitizes sensitive fields. The system's built-in data preprocessing module standardizes heterogeneous source data, including normalizing job titles (e.g., "VP" becomes "Vice President") and converting industry classification codes (referring to the GB / T 4754-2017 standard). When data conflicts are detected across multiple sources, the HR system master data is prioritized, triggering a log audit event. The user profile database utilizes a Redis cluster for millisecond-level real-time queries, and an LRU cache eviction mechanism ensures rapid response for frequently accessed data.

[0040] S220 : Match multiple candidate styles in a predetermined style library based on the current user's information.

[0041] Among them, the correspondence between multiple sample styles and user information can be determined in advance; each sample style and the corresponding user information are respectively input into the large language model to obtain a vector representation of each sample style and a vector representation of each user information; based on the vector representations of multiple sample styles, the vector representations of each user information and the correspondence between the sample style and the user information, a style library is determined.

[0042] The current user's information can be input into the large language model to obtain a vector representation of the current user's information; based on the vector representation of the current user's information, matching is performed in the style library to determine the sample styles corresponding to the closest multiple user information as multiple candidate styles.

[0043] In some embodiments, each sample style (such as "minimalist flat style") can be multimodally encoded. The visual encoder can extract visual feature vectors from design samples (such as UI interfaces), including key visual elements such as layout, color scheme, and font style. The output is a 512-dimensional dense vector. For example, the ViT model can be used to extract embedding vectors (dimension 512) for layout, color scheme, and font. The classification layer is removed, and the [CLS] tag output by the last layer of Transformer is extracted as the global visual embedding vector. The ViT (Vision Transformer) model can divide the input image into fixed-size image blocks (such as 16x16 pixels), convert them into embedding vectors through linear projection, and add position encoding to preserve space. It is also possible to capture global dependencies between different image blocks through multi-layer Transformer encoders, such as the coordination of color scheme and layout. Pre-trained ViT models (such as CLIP-ViT) can also be used to align visual and textual semantic spaces through comparative learning to enhance the semantic expression ability of visual features.

[0044] The information text encoder encodes the design description text (such as "suitable for the financial industry") into a semantic vector, capturing the text's style description and industry-required keywords. The design description text (such as "suitable for the financial industry") can be encoded using the RoBERTa model, with the [CLS] tag output by the last Transformer layer used as the global representation of the text. The text encoder encodes the design description text (such as "suitable for the financial industry") into a semantic vector, capturing the text's style description and industry-required keywords. An attention mechanism is used to associate style keywords (such as "minimalist"), industry attributes (such as "finance"), and visual features.

[0045] The fusion matrix is ​​responsible for mapping the visual and textual embedding vectors into a unified semantic space, enhancing cross-modal feature complementarity. The distribution of visual and textual vectors is adjusted through a learnable weight matrix (such as MLP or cross-attention). It is possible to combine early fusion (concatenating visual and textual vectors) with late fusion (weighted summation), optimizing the fusion weights through backpropagation. Fusion strategy: .in, Can be a fusion style vector, is a learnable fusion matrix that learns the weight distribution of visual and text features through back-propagation optimization. Assuming that the fusion style vector is 256-dimensional, then The dimension of is 1024×256. The fused It also preserves the semantic information of vision and text.

[0046] User profile vectors aggregate multidimensional user information (such as personal information, behavior sequences, and scenario context) to generate a 256-dimensional low-dimensional representation for personalized recommendations. The GraphSAGE model can be used to aggregate multidimensional user information (personal information nodes, behavior sequence edges, and scenario context nodes) to generate a low-dimensional dense vector (256 dimensions). GraphSAGE, based on the inductive learning model of graph neural networks, generates node embeddings by sampling neighboring nodes and aggregating their features (such as mean, LSTM, or attention). It then models user nodes (personal information), behavior edges (click sequences), and scenario nodes (usage environment) as a heterogeneous graph, capturing complex interactions through hierarchical aggregation.

[0047] A heterogeneous graph can be constructed with the user as the central node, connecting behavior logs (edges) and scene contexts (nodes).

[0048] Multi-hop neighbors can be randomly sampled (e.g., behavior sequences within 2 hops) to balance computational efficiency and information integrity.

[0049] Neighbor features can be fused using mean aggregation or attention mechanisms to optimize the model for link prediction or classification tasks.

[0050] S230: Taking the difference between the feature values ​​of the multiple candidate styles and the feature value of the current style as a style deviation.

[0051] The current style can be input into a large language model to obtain a vector representation of the current style. Based on the vector representation of the current style and the vector representations of multiple candidate styles, the style deviation is determined. The style feature value can be the vector representation of the style or other values ​​that can reflect the characteristics of the style or the differences between styles.

[0052] The sub-deviation between the vector representation of the current style and the vector representation of each alternative style can be determined based on the following formula:

[0053] ;

[0054] Among them, U is the vector representation of the current style;

[0055] Si is the vector representation of the i-th candidate style;

[0056] ; , the covariance matrix Σ is estimated by the sample covariance;

[0057] Used to constrain the matching of discrete features (such as font type);

[0058] Denominator Prevent similarity from being dominated by extreme values.

[0059] Among them, the style deviation is the difference between the mean of the eigenvalues ​​of multiple alternative styles and the eigenvalue of the current style; or, the style deviation includes the difference between the eigenvalues ​​of multiple alternative styles and the eigenvalue of the current style, that is, the style deviation can be a collection of multiple differences.

[0060] S240: Input the style deviation and the current business card into a generative large model to obtain one or more recommended business cards.

[0061] A prompt word template is predetermined, and the prompt word template includes a first content to be input and a second content to be input. The first content to be input is used to input a vector representation of the current style, and the second content to be input is used to input a style deviation.

[0062] The prompt template is as follows: "You are a senior graphic designer and need to generate a business card design based on the following constraints:

[0063] Must follow the current style features: [CLS] current vector representation [SEP];

[0064] The following deviations need to be balanced: [CLS] deviation vector description [SEP];

[0065] The output must meet the following requirements:

[0066] Brand specifications: ΔE of corporate VI color palette <5, font copyright compliance;

[0067] User preferences: elements that were frequently used in the past seven days (e.g., dark backgrounds accounted for 60%);

[0068] Scene adaptation: The current scene is [scene type];

[0069] [Generate Requirements];

[0070] Provide 3 design solutions, focusing on:

[0071] Option 1: Extreme professionalism (applicable to the financial / legal industries);

[0072] Option 2: Innovative visual expression (applicable to technology / creative industries);

[0073] Option 3: High information density (applicable to sales / consulting industries)".

[0074] S250: Display multiple recommended business card styles on a style change recommendation interface.

[0075] After the recommendation interface is displayed, it can respond to the style selection instructions triggered by the user on the recommendation interface, and perform dynamic rendering and multi-terminal synchronization of the business card style. In specific implementation, the rendering engine can use WebGL 2.0 to achieve GPU accelerated rendering, and implement anti-aliasing processing based on SDF signatures for vector elements (such as corporate logos). When the user clicks on a recommended style, the system calls the WASM-compiled image processing module to perform real-time style transfer, and uses the CycleGAN model to adapt the original business card visual elements to the new style while retaining the readability of the text. In terms of synchronization mechanism, a duplex communication channel is established through WebSocket, and the Operational Transformation algorithm is used to resolve multi-device editing conflicts. The style change records are persistently stored in the blockchain node in the JSON Patch format to ensure that the operation is traceable.

[0076] A closed loop of style library optimization driven by user feedback can also be built. An implicit feedback collection module is embedded in the recommendation interface: the user's dwell time on each recommended style (accurate to milliseconds), zoom operation trajectory (gesture path fitting through Bezier curves), and favorites addition behavior are recorded. Explicit feedback collects design satisfaction through a five-point Likert scale, and combines eye tracking data (pupil coordinates collected through the front camera) to calculate the visual focus heat map. After feature extraction, the feedback data is input into the reinforcement learning model (PPO algorithm) to dynamically adjust the weight parameters in the style matching strategy. Incremental updates to the style library are performed regularly every week: styles with zero exposure for 30 days are eliminated, and data enhancement is performed on high-conversion-rate styles (including 12 image transformations such as hue shift of ±15° and kerning fine-tuning of ±2%). New styles must pass the aesthetic evaluation model (based on the NIMA architecture) and score more than 8.5 points before they can be added to the library.

[0077] It can also ensure cross-media style compatibility. When it is detected that the user has enabled the AR business card function, the 3D style adaptation pipeline is automatically activated: the 3D reconstruction model of the business card is generated through NeRF technology, and the material properties are recalculated using the physical rendering (PBR) workflow. For printed scenes, the color management module is called to perform CMYK color separation calibration, and G7 certified color matching is implemented based on the ICC profile. 1mm bleed and 300dpi printing pre-processing files are generated simultaneously. The responsive layout engine monitors the DPI parameters of the display medium in real time. When switching to small-screen devices such as smart watches, it automatically triggers information hierarchy reconstruction - the core information (name / position) is identified through the CRF conditional random field model, and the secondary information (company introduction QR code) is progressively folded.

[0078] In some embodiments, a style matching matrix can be constructed based on multi-dimensional features to generate a personalized recommendation candidate set. A feature fusion layer concatenates the layout feature vector (0.78, 0.12, 0.65), color histogram (#3F88C5 dominant color accounts for 32%), and texture complexity coefficient (DCT frequency domain energy 0.45) with the user feature vector. This is then fed into a deep cross-validation network (DCN) to calculate a style fit score. By leveraging the responsive grid parameters of enterprise-level design systems (such as Ant Design) and combining them with Material Design's spacing and ratio specifications, an accessible color scheme that complies with WCAG 2.1 AA standards is dynamically generated. For users in the financial industry, the visual weight of security elements (such as the shield icon) is automatically enhanced, while for users in the technology sector, the placeholder area for the dynamic data visualization module is increased.

[0079] In some embodiments, a real-time feedback closed-loop optimization recommendation strategy can also be established. When a user selects a new style, implicit feedback data such as dwell time (≥5s triggers a valid event) and edit rollback rate (number of undo operations) are collected. The user behavior event time window is constructed through the Flink stream processing engine, and the user preference model is updated every 5 minutes. When designing the multi-objective optimization function, the compliance with the corporate VI specification (weight 0.6) and the personalized innovation index (weight 0.4) are balanced, and the NSGA-II algorithm is used to solve the Pareto optimal solution set. For users who make frequent modifications (≥3 operations per day), the reinforcement learning mechanism is activated to dynamically adjust the exploration / utilization ratio, and the Q-learning algorithm is used to explore potential preference combinations in the style library.

[0080] In some embodiments, cross-terminal style synchronization and version control can also be implemented. When the mobile terminal completes the style change, it is synchronized to the PC-side corporate email signature through a WebSocket long connection, and the Jira API is called to update the user identity in the collaboration tool. The Git version management solution is used to record the style modification history, and supports difference comparison and rollback operations by timestamp (accurate to milliseconds) or semantic tags (such as "Business Official Version_v2.1.3"). When deeply integrated with WeChat for Business, a style change description card is automatically generated, and the modification details are rendered using Markdown syntax (adjusting the font spacing from 1.5 to 1.8, and the hue shift by +15°), and @ the relevant reviewers are confirmed for compliance.

[0081] In order to execute the corresponding steps in the above embodiments and various possible methods, the following is an implementation method of a device for changing the business card style with one click based on AI. Optionally, the device for changing the business card style with one click based on AI can adopt the above Figure 1 For further information, please refer to Figure 3 , Figure 3 This is a functional module diagram of an AI-based one-click business card style changing device provided by an embodiment of the present invention. It should be noted that the basic principle and technical effects of the AI-based one-click business card style changing device provided by this embodiment are the same as those of the above embodiments. For the sake of brief description, for parts not mentioned in this embodiment, please refer to the corresponding content in the above embodiments. The AI-based one-click business card style changing device includes:

[0082] The receiving module 301 is configured to receive a one-key change of business card style operation triggered by a user for a current business card, and determine the current style of the current business card and information about the current user;

[0083] A matching module 302 is configured to match a plurality of candidate styles in a predetermined style library based on information of a current user;

[0084] a determination module 303 for using the difference between the characteristic values ​​of the multiple candidate patterns and the characteristic value of the current pattern as a pattern deviation;

[0085] A generation module 304 is used to input the style deviation and the current business card into a generative model to obtain one or more recommended business cards;

[0086] The recommendation module 305 is used to display a plurality of recommended business cards in the style change recommendation interface.

[0087] In some embodiments, the style of the business card includes font, color, and layout; the user information includes user personal information, user behavior information, and user scene information.

[0088] In some embodiments, the pattern deviation is the difference between the mean of the feature values ​​of the multiple candidate patterns and the feature value of the current pattern; or, the pattern deviation includes the difference between the feature values ​​of the multiple candidate patterns and the feature value of the current pattern.

[0089] In some embodiments, the device also includes an establishment module for: pre-determining the correspondence between multiple sample styles and user information; inputting each sample style and the corresponding user information into a large language model to obtain a vector representation of each sample style and a vector representation of each user information; wherein the characteristic value of the style is the vector representation of the style; and determining a style library based on the vector representation of multiple sample styles, the vector representation of each user information, and the correspondence between the sample style and the user information.

[0090] In some embodiments, the matching module 302 is specifically used to: input the current user's information into the large language model to obtain a vector representation of the current user's information; perform matching in the style library based on the vector representation of the current user's information, and determine the sample styles corresponding to the closest multiple user information as multiple alternative styles.

[0091] In some embodiments, the determination module 303 is specifically configured to: input the current style into the large language model to obtain a vector representation of the current style; and determine a style deviation based on the vector representation of the current style and vector representations of multiple candidate styles.

[0092] In some embodiments, a prompt word template is predetermined, and the prompt word template includes a first content to be input and a second content to be input. The first content to be input is used to input a vector representation of the current style, and the second content to be input is used to input a style deviation.

[0093] Optionally, the above modules can be stored in the form of software or firmware. Figure 1 The memory shown in the figure may be solidified in the operating system (OS) of the AI-based one-key business card style changer and can be Figure 1 Meanwhile, the data and program codes required to execute the above modules may be stored in the memory.

[0094] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a portion of code, and the module, program segment or a portion of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or can be implemented using a combination of dedicated hardware and computer instructions.

[0095] In addition, the functional modules in the various embodiments of the present invention may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0096] If the functions are implemented in the form of software modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, electronic device, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0097] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for changing business card styles with one click based on AI, characterized in that: include: Receive a one-click change of business card style operation triggered by the user for the current business card, and determine the current style of the current business card and the current user's information; Inputting the current user's information into a large language model to obtain a vector representation of the current user's information; Based on the vector representation of the current user's information, matching is performed in a style library to determine sample styles corresponding to the closest plurality of user information as a plurality of candidate styles; wherein the correspondence between the plurality of sample styles and the user information is predetermined; each of the sample styles and the corresponding user information is input into a large language model to obtain a vector representation of each sample style and a vector representation of each user information; wherein the feature value of the style is the vector representation of the style; and a style library is determined based on the vector representations of the plurality of sample styles, the vector representations of each user information, and the correspondence between the sample styles and the user information; taking the difference between the feature values ​​of the plurality of candidate patterns and the feature value of the current pattern as a pattern deviation; Inputting the style deviation and the current business card into a generative large model to obtain one or more recommended business card styles; wherein a prompt word template is predetermined, the prompt word template including first content to be input and second content to be input, the first content to be input being used to input a vector representation of the current style, and the second content to be input being used to input the style deviation; the generative large model, based on the prompt word template and following the vector representation of the current style, balances the style deviation to obtain one or more recommended business card styles; Display multiple recommended business card styles on the style change recommendation interface; After the recommendation interface is displayed, in response to a style selection instruction triggered by the user on the recommendation interface, dynamic rendering of the business card style and multi-terminal synchronization are performed; The style of the business card includes font, color and layout; the user information includes user personal information, user behavior information and user scene information; Determining the style deviation based on the multiple candidate styles and the current style includes: inputting the current style into a large language model to obtain a vector representation of the current style; determining the style deviation based on the vector representation of the current style and the vector representations of the multiple candidate styles; and determining a sub-deviation between the vector representation of the current style and the vector representation of each candidate style based on the following formula: ； Among them, U is the vector representation of the current style; Si is the vector representation of the i-th candidate style; ; , the covariance matrix Σ is estimated by the sample covariance; Used to constrain the matching of discrete features; Denominator Used to prevent similarity from being dominated by extreme values.

2. The method according to claim 1, characterized in that The pattern deviation is the difference between the mean of the feature values ​​of the multiple candidate patterns and the feature value of the current pattern; Alternatively, the pattern deviation includes differences between feature values ​​of a plurality of candidate patterns and feature values ​​of the current pattern.

3. An AI-based device for changing business card styles with one click, characterized in that: include: A receiving module is used to receive a one-key change of business card style operation triggered by a user for the current business card, and determine the current style of the current business card and the information of the current user; a matching module, configured to input the current user's information into a large language model to obtain a vector representation of the current user's information; Based on the vector representation of the current user's information, matching is performed in a style library to determine sample styles corresponding to the closest plurality of user information as a plurality of candidate styles; wherein the correspondence between the plurality of sample styles and the user information is predetermined; each of the sample styles and the corresponding user information is input into a large language model to obtain a vector representation of each sample style and a vector representation of each user information; wherein the feature value of the style is the vector representation of the style; and a style library is determined based on the vector representations of the plurality of sample styles, the vector representations of each user information, and the correspondence between the sample styles and the user information; a determination module, configured to use a difference between the characteristic values ​​of the plurality of candidate patterns and the characteristic value of the current pattern as a pattern deviation; A generation module is configured to input the style deviation and the current business card into a generative large model to obtain one or more recommended business card styles. A prompt word template is predetermined, and the prompt word template includes first content to be input and second content to be input, the first content to be input being used to input a vector representation of the current style, and the second content to be input being used to input the style deviation. The generative large model, based on the prompt word template and following the vector representation of the current style, balances the style deviation to obtain one or more recommended business card styles. A recommendation module is used to display multiple recommended business card styles on a style change recommendation interface; after the recommendation interface is displayed, in response to a style selection instruction triggered by the user on the recommendation interface, perform dynamic rendering of the business card style and synchronize it with multiple terminals; The style of the business card includes font, color and layout; the user information includes user personal information, user behavior information and user scene information; The matching module is further configured to input the current style into a large language model to obtain a vector representation of the current style; determine a style deviation based on the vector representation of the current style and the vector representations of the multiple candidate styles; and determine a sub-deviation between the vector representation of the current style and the vector representation of each candidate style based on the following formula: ； Among them, U is the vector representation of the current style; Si is the vector representation of the i-th candidate style; ; , the covariance matrix Σ is estimated by the sample covariance; Used to constrain the matching of discrete features; Denominator Used to prevent similarity from being dominated by extreme values.

4. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement the method according to any one of claims 1 to 2.

Images