Advertisement cover generation method and device and storage medium

Abstract

The embodiment of the invention provides an advertisement cover generation method and device and a storage medium, and the method comprises the steps: obtaining user portrait data corresponding to a client and advertisement cover template data corresponding to a to-be-delivered advertisement if it is monitored that the client starts a target application program; according to the user portrait data and the advertisement cover template data, determining prompt words corresponding to content elements in a to-be-generated target advertisement cover; fusing the cue words corresponding to the content elements to form a target cue word; and inputting the target prompt word into an advertisement cover generation model, generating a target advertisement cover corresponding to the client, and displaying the target advertisement cover in a target application program corresponding to the client. According to the scheme, the cue words corresponding to the content elements are generated according to the user portrait data and the advertisement cover template data, and the fused cue words are input into the advertisement cover generation model, so that the personalized advertisement cover corresponding to the client can be accurately generated.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an advertisement cover generation method, device and storage medium. Background Art

[0002] With the development of the Internet and mobile terminals, the public's attention has shifted to mobile terminals. Currently, a large number of advertisements have begun to be placed in third-party applications on such mobile devices. However, the advertisements placed in various applications are currently designed by advertisers. Therefore, when each user opens the application, the advertisement cover displayed to the user is the same, and it is impossible to display it differently for different users, which reduces the attractiveness of the advertisement. Summary of the invention

[0003] The embodiment of the present invention provides an advertisement cover generation method, device and storage medium, which are used to customize and generate personalized advertisement covers corresponding to each client to achieve differentiated display.

[0004] In a first aspect, an embodiment of the present invention provides an advertisement cover method, the method comprising:

[0005] If it is detected that the client starts the target application, the user portrait data corresponding to the client and the advertisement cover template data corresponding to the advertisement to be placed are obtained;

[0006] Determine, according to the user portrait data and the advertisement cover template data, prompt words corresponding to each content element in the target advertisement cover to be generated;

[0007] Merging the prompt words corresponding to the various content elements to integrate the prompt words corresponding to the various content elements into a target prompt word;

[0008] Inputting the target prompt words into the advertisement cover generation model to generate target content corresponding to each content element;

[0009] According to the target content corresponding to each content element, a target advertisement cover corresponding to the client is generated, so as to display the target advertisement cover in a target application corresponding to the client.

[0010] In a second aspect, an embodiment of the present invention provides an advertisement cover generation device, the device comprising:

[0011] An acquisition module, for acquiring user portrait data corresponding to the client and advertisement cover template data corresponding to the advertisement to be placed if it is detected that the client starts the target application;

[0012] A determination module, used to determine the prompt words corresponding to each content element in the target advertisement cover to be generated according to the user portrait data and the advertisement cover template data;

[0013] A fusion module, used for fusing the prompt words corresponding to the various content elements, so as to integrate the prompt words corresponding to the various content elements into a target prompt word;

[0014] A first generating module, used for inputting the target prompt words into an advertisement cover generating model to generate target contents corresponding to the respective content elements;

[0015] The second generating module is used to generate a target advertisement cover corresponding to the client according to the target content corresponding to each content element, so as to display the target advertisement cover in a target application corresponding to the client.

[0016] In a third aspect, an embodiment of the present invention provides an electronic device, comprising: a memory, a processor, and a communication interface; wherein the memory stores executable code, and when the executable code is executed by the processor, the processor executes the advertising cover generation method as described in the first aspect.

[0017] In a fourth aspect, an embodiment of the present invention provides a non-temporary machine-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor can at least implement the advertising cover generation method as described in the first aspect.

[0018] In a fifth aspect, an embodiment of the present invention provides a computer program product, comprising: a computer program, when the computer program is executed by a processor of an electronic device, the processor executes the advertisement cover generation method as described in the first aspect.

[0019] In the advertising cover generation scheme provided by the embodiment of the present invention, when an advertisement is placed for the client during the period when the client uses the application, a personalized advertising cover is customized for the client to better suit the user's browsing interests. Specifically, if it is monitored that the client starts the target application, the user portrait data corresponding to the client and the advertising cover template data corresponding to the advertisement to be placed are obtained. Based on the user portrait data and the advertising cover template data, the prompt words corresponding to each content element in the target advertising cover to be generated are determined. The prompt words corresponding to each content element are merged to integrate multiple prompt words into one target prompt word. The target prompt word is input into the advertising cover generation model to generate the target content corresponding to each content element. Finally, based on the target content corresponding to each content element, the target advertising cover corresponding to the client is generated to display the target advertising cover in the target application corresponding to the client.

[0020] In the above scheme, the prompt words corresponding to each content element in the target advertisement cover to be generated are determined according to the user portrait data and the advertisement cover template data, and the prompt words corresponding to multiple content elements are integrated to guide the advertisement cover generation model to control the generation of each content element in the target advertisement cover finally generated according to the prompt words corresponding to the multiple content elements. This can not only improve the accuracy of the advertisement cover design, but also make the generated target advertisement cover highly relevant to the user's interests and needs, so as to customize the personalized advertisement cover corresponding to the client, thereby enhancing the advertising effect and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 A schematic diagram of an application scenario of a method for generating an advertisement cover provided by an embodiment of the present invention;

[0023] Figure 2 A flowchart of a method for generating an advertisement cover provided by an embodiment of the present invention;

[0024] Figure 3 A flowchart of another method for generating an advertisement cover provided by an embodiment of the present invention;

[0025] Figure 4 A flowchart for determining prompt words corresponding to each content element in a target advertisement cover to be generated provided by an embodiment of the present invention;

[0026] Figure 5 A schematic diagram of another application of generating an advertisement cover provided by an embodiment of the present invention;

[0027] Figure 6 A schematic diagram of the structure of an advertisement cover generation device provided by an embodiment of the present invention;

[0028] Figure 7 A schematic diagram of the structure of an electronic device provided in this embodiment. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In addition, the step timing in the following method embodiments is only an example, not a strict limitation.

[0030] It should be noted that the user data (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present invention are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0031] With the development of the Internet and mobile terminals, the public's attention has shifted to mobile terminals. Currently, a large number of advertisements have begun to be placed in third-party applications in such mobile devices. Among them, they are usually designed and pushed by advertisers. The ad cover is an important part of the advertisement. It is directly facing the audience and is the key to attracting the audience's attention. However, when each different user opens the application, the ad cover displayed to the user is the same for different users, and it is impossible to display it differently for different users, which reduces the attractiveness of the advertisement. In other words, the same advertisement is displayed in each client with a unified ad cover, and it is impossible to display it differently for different users.

[0032] Therefore, in order to solve the above technical problems, the present invention provides an advertising cover generation scheme, in which prompt words corresponding to each content element in the advertising cover are generated based on user portrait data and advertising cover template data corresponding to the advertisement to be placed, and the prompt words corresponding to each content element are integrated to better guide the advertising cover generation model to accurately control the generation of each content element according to the prompt words corresponding to each content element, so that each content element in the generated target advertising cover is accurately presented according to the design effect corresponding to the prompt word. This can not only ensure that the generated target advertising cover is highly relevant to the user's interests and needs, so as to customize the personalized advertising cover corresponding to the client, but also ensure the quality of the generated target advertising cover.

[0033] In the following, some embodiments of the present invention are described in detail in conjunction with the accompanying drawings. In the case where there is no conflict between the embodiments, the following embodiments and features in the embodiments can be combined with each other.

[0034] Figure 1 A schematic diagram of an application scenario of an advertisement cover generation method provided by an embodiment of the present invention; Figure 1 As shown, the execution subject of the method may be an advertisement cover generation device, which may be implemented as software, or a combination of software and hardware, and the advertisement cover generation device may be communicatively connected with a client.

[0035] The client may refer to a user device with a target application installed. The target application may be any type of application, and the embodiments of the present invention do not limit this. The advertisement cover generation device may automatically generate a target advertisement cover corresponding to the advertisement to be placed that matches the client on the client, and dynamically adjust the advertisement cover corresponding to the advertisement to be placed in the target application in the client according to the actual needs and interests of the user, so as to achieve more accurate placement of the advertisement cover that the user is interested in, which can not only improve the advertisement conversion rate but also enhance the user experience.

[0036] Among them, the advertisement cover generation device can regularly collect basic user data, user behavior data, and user feedback data on displayed advertisement covers corresponding to the client. When the client is using the target application, the user portrait data corresponding to the client can be generated based on the various information collected corresponding to the client, so as to have a deeper understanding of the user's interests, preferences, behavior patterns, etc. The collected user data and user portrait data can also be stored in a preset storage area, so that the user portrait data corresponding to the client can be directly obtained from the preset storage area later.

[0037] Furthermore, the advertisement cover generation device can also periodically collect advertisements to be placed, and store the advertisement cover templates corresponding to the data to be placed in a preset storage area. In addition, in actual applications, when it is detected that the user starts the target application, the basic user data, user behavior data, user feedback data corresponding to the client, and the advertisement cover templates corresponding to the advertisements to be placed can be collected.

[0038] After completing the above processing, the startup status of the target application in the client is monitored. If the client is detected to have started the target application, a target advertising cover that meets the user's interests or needs is generated for the client. In this way, when the user starts the target application through the client, the target advertising cover that matches the user's interests and hobbies can be automatically displayed in the target application while the user is using the target application. This eliminates the need for manual operation by advertising operators and is more intelligent and efficient.

[0039] Among them, the specific implementation process of generating a target advertisement cover for a client includes: if it is monitored that the client starts a target application, the user portrait data corresponding to the client and the advertisement cover template data corresponding to the advertisement to be placed are obtained. According to the user portrait data and the advertisement cover template data, the prompt words corresponding to each content element in the target advertisement cover to be generated are determined. The prompt words corresponding to each content element are merged to integrate multiple prompt words into one target prompt word. The merged target prompt words are input into the advertisement cover generation model to generate the target content corresponding to each content element. According to the target content corresponding to each content element, the target advertisement cover corresponding to the client is generated to display the target advertisement cover in the target application corresponding to the client.

[0040] The technical solution provided in this embodiment generates prompt words corresponding to each content element in the advertisement cover based on user portrait data and advertisement cover template data corresponding to the advertisement to be placed, and merges the prompt words corresponding to each content element to better guide the advertisement cover generation model to accurately control the generation of each content element according to the prompt words corresponding to each content element, so as to customize the personalized advertisement cover corresponding to the client.

[0041] The above embodiment describes the advertisement cover generation process in a specific application scenario. In order to more clearly understand the above advertisement cover generation process, the generation process is described in detail in combination with the following embodiment.

[0042] Figure 2 A flowchart of a method for generating an advertisement cover provided by an embodiment of the present invention is shown in FIG. Figure 2 As shown, the method comprises the following steps:

[0043] 201. If it is detected that the client starts the target application, the user portrait data corresponding to the client and the advertisement cover template data corresponding to the advertisement to be placed are obtained.

[0044] 202. Determine prompt words corresponding to various content elements in the target advertisement cover to be generated according to the user portrait data and the advertisement cover template data;

[0045] 203. Merge the prompt words corresponding to the various content elements to integrate the prompt words corresponding to multiple content elements into one target prompt word.

[0046] 204. Input the target prompt words into the advertisement cover generation model to generate target content corresponding to each content element.

[0047] 205. Generate a target advertisement cover corresponding to the client according to the target content corresponding to each content element, so as to display the target advertisement cover in a target application corresponding to the client.

[0048] In specific implementation, it is possible to monitor whether the target application in the client is started. If it is detected that the client has started the target application, the user portrait data corresponding to the client and the advertisement cover template data corresponding to the advertisement to be placed are obtained. Among them, the user portrait data of the client can comprehensively reflect the characteristics of the user from multiple dimensions, and provide strong data support for the user to generate a target advertisement cover that is more interesting. The advertisement cover template data may include information such as advertisement cover size, design style, font selection, copywriting, images, icons, etc.

[0049] Optionally, the specific implementation process of obtaining the user portrait data corresponding to the client can be: collecting the user data corresponding to the client. Among them, the user data includes basic user information, user behavior data, etc. Basic information, such as age, gender, geographic location, occupation, etc. User behavior data, such as frequency of use, access time, interaction method, etc. Then, the collected user data is pre-processed by cleaning, deduplication, format conversion, etc. to ensure the accuracy and availability of the data. Furthermore, natural language NLP technology is used to analyze user behavior data such as search history, browsing content, interactive behavior, etc. in the target application. Obtain user interest preferences, user behavior patterns (for example, browsing habits, preferred browsing style), etc. Then, combined with user basic information, user interest preferences and user behavior patterns, user portrait data is constructed through a deep learning algorithm.

[0050] Among them, the specific implementation process of obtaining user interest preferences and user behavior patterns can be: parsing the search keywords entered by the user through NLP technology to understand their semantic intent. Extracting keywords and performing frequency statistics to identify the topics or fields that users frequently search. Analyzing the time series of search history to identify the changing trend of user interests. In addition, using NLP technology to perform text analysis on the page content browsed by the user to extract key information and topics. Using sentiment analysis technology to evaluate the user's emotional tendency towards the browsed content. Analyzing the user's dwell time and click behavior on different types of content to identify the user's preferred browsing style. At the same time, monitoring the user's interactive behaviors such as likes, comments, and sharing in the target application. Analyzing the semantic features of the user's interactive content through NLP technology to identify the user's interests and opinions. Analyzing the frequency and pattern of interactive behaviors to identify user behavior habits. Finally, using the DL deep learning algorithm, the user's search topics or fields, the changing trend of user interests, the emotional tendency towards the browsed content, the preferred browsing style, the user's interests and opinions, the user's behavior habits, and the user's basic information are analyzed to construct user portrait data.

[0051] After obtaining the user portrait data corresponding to the client, the advertisement cover template corresponding to the advertisement to be placed is then obtained, and the advertisement cover template is parsed to obtain the advertisement cover template data corresponding to the advertisement to be placed. Among them, when parsing the advertisement cover template, key information and design elements in the advertisement cover template can be extracted. Among them, key information may include advertisement title, cover image, promotional copy, advertisement image, etc. Design elements are elements used to construct the visual effect of the advertisement cover, and they create unique visual effects through specific combinations and arrangements. Its design elements include background color, text color, layout, font, etc.

[0052] Next, we use generative artificial intelligence (AIGC) technology to combine user portrait data and advertising cover template data to generate the target advertising cover.

[0053] In order to more accurately control the generation of the target ad cover, to ensure that the generated target ad cover is highly relevant to the user's interests and needs, and to ensure that the generated target ad cover retains the key information in the original ad cover template, the prompt words corresponding to each content element in the target ad cover to be generated can be determined based on the user portrait data and the ad cover template data corresponding to the ad to be placed. Then use the generated multiple prompt words to guide the ad cover generation model, and generate the target ad cover according to the prompt words corresponding to each content element. Among them, content elements refer to elements in the ad cover that are used to convey specific information and themes. Content elements can include ad titles, cover images, promotional copy, ad images, etc.

[0054] Among them, in order to better control the various content elements in the generated target advertising cover, when generating prompt words, the various content elements in the target advertising cover can be determined based on the advertising cover template data, and then the description information corresponding to each content element can be determined in combination with the user portrait data. Based on the description information corresponding to each content element, the prompt words corresponding to each content element are generated in turn.

[0055] From the above description, we can see that by carefully designing the prompt words corresponding to each content element in the target advertisement cover to be generated, we can guide the advertisement cover generation model to accurately understand the task requirements, and accurately control the generation of each content element according to the prompt words corresponding to each content element, thereby improving the efficiency and accuracy of task processing.

[0056] In addition, the specific implementation method for determining the prompt words corresponding to each content element in the target advertisement cover to be generated is not limited to this. In the embodiment of the present invention, the user portrait data and the advertisement cover template data can be automatically matched based on the preset prompt word generation rules or preset logical conditions, and the prompt words corresponding to each content element can be generated. Alternatively, the user portrait data and the advertisement cover template data can be processed using a machine learning model to generate prompt words corresponding to each content element. Among them, the machine learning model is used to identify the relationship between the user portrait data and the best advertisement element.

[0057] After generating the prompt words corresponding to each content element, the prompt words corresponding to each content element are then fused to integrate multiple prompt words into a target prompt word. The target advertising cover corresponding to the client is generated by combining the fused target prompt words through AIGC artificial intelligence content generation technology, so as to display the target advertising cover in the target application corresponding to the client.

[0058] Prompt word fusion refers to combining multiple prompt words into a text description to guide the model to produce more accurate, useful or expected target content. The structure, order or format of the prompt words can be adjusted to ensure that the target prompt words generated in the end can guide the AI ​​model to produce the expected output.

[0059] Specifically, the prompt words corresponding to each content element can be directly listed together to integrate multiple prompt words into a target prompt word; or the prompt words corresponding to multiple content elements can be arranged in sequence according to the logical order or importance, and the multiple prompt words can be combined in sequence according to the arrangement order to integrate multiple prompt words into a target prompt word; or the key information in the prompt words corresponding to multiple content elements can be integrated into a sentence or paragraph according to a preset prompt word template to integrate multiple prompt words into a target prompt word. Or the differences between the prompt words corresponding to multiple content elements can be compared, and then the fused target prompt word can be generated according to the difference information between the multiple prompt words.

[0060] The target prompt words generated in this way can guide the advertising cover generation model to better understand the specific requirements and context of the task. In addition, in addition to the target content corresponding to multiple content elements, the target prompt words can also include defined tasks to be processed, set context information, and the format corresponding to the output results, etc., to ensure that the content output by the model meets the task requirements.

[0061] That is to say, all prompt words are integrated into a coherent text description to ensure that this description can accurately convey the overall information and style of the ad cover. Natural language processing (NLP) technology can be used to optimize the connection between prompt words corresponding to multiple content elements to make them more natural and smooth, thereby ensuring the effect and quality of the target ad cover generated by AIGC artificial intelligence content generation technology.

[0062] Specifically, the fused target prompt words can be input into the advertisement cover generation model to generate target content corresponding to each content element. The advertisement cover generation model is pre-trained to generate target content corresponding to each content element in the advertisement cover.

[0063] Finally, according to the target content corresponding to each content element, a target advertisement cover corresponding to the client is generated, so as to display the target advertisement cover in the target application corresponding to the client.

[0064] In addition, the present embodiment does not limit the specific implementation method for generating the target advertisement cover, and those skilled in the art may set it according to specific application requirements. For example, a matching target advertisement cover template may be screened out from a plurality of preset advertisement cover templates, and the target contents corresponding to the plurality of content elements may be filled into the corresponding positions of the target advertisement cover template to obtain the target advertisement cover. Alternatively, preset rules may be used to combine the target contents corresponding to the plurality of content elements to generate the target advertisement cover. Alternatively, a pre-trained machine learning model may be used to predict the combination method corresponding to the plurality of content elements, and the target contents corresponding to the plurality of content elements may be combined using the machine learning model to generate the target advertisement cover.

[0065] In addition, when combining multiple content elements, you can first optimize the target content corresponding to the multiple content elements based on preset content generation rules, and then combine the target content corresponding to the optimized multiple content elements, so that the generated target advertising cover is more beautiful and harmonious.

[0066] The embodiment of the present invention determines the prompt words corresponding to each content element in the target advertisement cover to be generated according to user portrait data and advertisement cover template data, and integrates the prompt words corresponding to multiple content elements to guide the advertisement cover generation model to control the generation of each content element in the target advertisement cover finally generated according to the prompt words corresponding to the multiple content elements. This can not only improve the accuracy of the advertisement cover design, but also make the generated target advertisement cover highly relevant to the user's interests and needs, so as to customize the personalized advertisement cover corresponding to the client, thereby enhancing the advertising effect and user experience.

[0067] The above embodiment describes the generation of prompt words corresponding to each content element based on user portrait data and advertisement cover template data, so as to guide the advertisement cover generation model, accurately control the generation of each content element, and then generate a target advertisement cover according to the target content corresponding to each content element. Among them, in order to make the generated target advertisement cover more beautiful and more attractive to the client, when generating the target advertisement cover, multiple design elements corresponding to the target advertisement cover and prompt words corresponding to multiple design elements can also be determined at the same time, so that the style, text and image combination method, hierarchical expression, layout, format, etc. corresponding to the generated multiple content elements can also be accurately controlled.

[0068] The specific implementation process is described in detail below in conjunction with the following embodiments.

[0069] Figure 3 A flowchart of another method for generating an advertisement cover provided by an embodiment of the present invention is shown in FIG. Figure 3 As shown, the method comprises the following steps:

[0070] 301. If it is detected that the client starts the target application, the user portrait data corresponding to the client and the advertisement cover template data corresponding to the advertisement to be placed are obtained.

[0071] 302. Determine prompt words corresponding to various content elements in the target advertisement cover to be generated based on the user portrait data and the advertisement cover template data.

[0072] 303. Determine multiple design elements corresponding to the target advertisement cover.

[0073] 304. Determine display information corresponding to multiple design elements based on the user portrait data.

[0074] 305. Generate prompt words corresponding to the multiple design elements according to the display information corresponding to the multiple design elements.

[0075] 306. Merge the prompt words corresponding to each content element and the prompt words corresponding to the multiple design elements to integrate the prompt words corresponding to each content element and the prompt words corresponding to the multiple design elements into a target prompt word.

[0076] 307. Input the target prompt words into the advertisement cover generation model to generate target content corresponding to each content element.

[0077] 308. Generate a target advertisement cover corresponding to the client according to the target content corresponding to each content element, so as to display the target advertisement cover in a target application corresponding to the client.

[0078] The description of step 301 and step 302 is referred to the above embodiment, and will not be repeated here. After determining the prompt words corresponding to each content element in the target advertisement cover to be generated, then determine multiple design elements corresponding to the target advertisement cover. The design elements include fonts, colors, layouts, etc.

[0079] In specific implementation, multiple design elements corresponding to the target ad cover can be determined based on the ad cover template data and the ad cover generation rules corresponding to the ad to be placed. The ad cover generation rules refer to a series of guiding principles and specific standards followed when using automated tools or AI models to create ad covers. These rules are designed to ensure that the generated ad cover is not only beautiful and attractive, but also effectively conveys brand information, product features, and key points of marketing activities.

[0080] Among them, the ad cover generation rules can be generated in advance according to the needs of advertisers, advertising scenarios, advertising subjects, content guidelines, privacy and data security. For example, the ad cover generation rules specify the key elements of the advertisement contained in the target ad cover, such as stipulating that the ad cover content includes the title, subtitle, promotional copy, background image, ad image, brand logo, platform logo, layout design, etc. Alternatively, the ad cover generation rules specify the specific environment and background set by the ad content, which helps to determine the target audience and the message conveyed by the ad. For example: online shopping, social media, content marketing, outdoor advertising, application promotion, etc. The ad cover generation rules specify the core information and the main purpose of the ad content, which directly affects the attractiveness and effect of the ad. For example: new product launch, seasonal promotion, brand image enhancement, etc. Alternatively, the ad cover generation rules specify some guiding principles for personalized content, such as: information accuracy, copywriting style, visual style, etc. Alternatively, the ad cover generation rules specify requirements for privacy compliance and data security.

[0081] Among them, the advertisement cover generation rules may include at least one of the above rules, or may include all of the above rules, and there is no limitation on this.

[0082] After determining the multiple design elements corresponding to the target advertisement cover, the display information corresponding to the multiple design elements is determined based on the user portrait data, so as to ensure that the determined design elements can better meet the user preferences. In specific implementation, a large amount of historical data can be collected, and the preference patterns corresponding to each user group can be determined using a clustering algorithm. Then, the client can be classified using a classification algorithm to determine the type of user group corresponding to the client. Furthermore, based on the characteristics of the user group corresponding to the client, the type of design that the client may like is determined, and then the display information corresponding to the multiple design elements is determined. Among them, the distance algorithm can be K-means, DBSCAN, etc. The classification algorithm can be a decision tree, random forest, etc.

[0083] Next, based on the display information corresponding to the multiple design elements, prompt words corresponding to the multiple design elements are generated. That is, the display information corresponding to the multiple design elements is converted into a language that can be recognized by the advertisement cover generation model, so as to accurately control the style and pattern corresponding to each content element generated by the advertisement cover generation model.

[0084] In actual applications, there are conflicts between the display information corresponding to the determined design elements. In order to obtain a higher quality target advertisement cover, in an optional embodiment, after the display information corresponding to multiple design elements is determined, the display information corresponding to the multiple design elements can be optimized by simulating natural selection and genetic mechanisms to obtain the optimized target display information corresponding to the multiple design elements; and prompt words corresponding to the multiple design elements are generated according to the target display information corresponding to the multiple design elements.

[0085] Specifically, create an initial set of design elements, each member of which represents a possible design solution. Convert each design solution into a form that can be processed by a computer, such as using binary strings, real number vectors, or character strings to represent different design elements and their attributes. Set a fitness function to evaluate the quality of each design solution. This can be based on preset aesthetic standards, user feedback, click-through rate, and other indicators. Then, select the design solution with better performance as the parent individual according to the fitness value. Randomly select two individuals from the selected parent generation and exchange part of their codes to generate new child individuals. Make random changes with a small probability to the newly generated offspring to introduce diversity. Repeat the selection, crossover, and mutation process for multiple rounds until the stopping condition is met, so that the target display information corresponding to the optimized multiple design elements can be determined.

[0086] Next, the prompt words corresponding to each content element and the prompt words corresponding to multiple design elements are merged to integrate the prompt words corresponding to multiple content elements and the prompt words corresponding to multiple design elements into a target prompt word. The merged target prompt word is input into the advertisement cover generation model to generate the target content corresponding to each content element, and the target advertisement cover corresponding to the client is generated according to the target content corresponding to each content element, so as to display the target advertisement cover in the target application corresponding to the client.

[0087] In summary, the embodiments of the present invention determine the prompt words corresponding to each content element in the target advertisement cover according to user portrait data and advertisement cover template data, determine the prompt words corresponding to multiple design elements according to the user portrait data, and merge the prompt words corresponding to multiple content elements and multiple design elements to guide the advertisement cover generation model according to the prompt words corresponding to multiple content elements and the prompt words corresponding to multiple design elements. This can not only accurately control the specific content as well as the specific style and style corresponding to each content element in the finally generated target advertisement cover, but also customize a higher quality target advertisement cover for the client.

[0088] In conjunction with the following embodiments, the specific implementation process of determining the prompt words corresponding to each content element in the target advertisement cover to be generated based on the user portrait data and the advertisement cover template data in the above embodiment is described in detail.

[0089] Figure 4 A flowchart of determining prompt words corresponding to each content element in a target advertisement cover to be generated is provided in an embodiment of the present invention; based on the above embodiment, refer to the attached Figure 3 As shown, the method comprises the following steps:

[0090] 401. Obtain the user feedback data of the client on the historical advertisements that have been displayed.

[0091] 402. Determine the user preference characteristics corresponding to the client according to the user portrait data and the user feedback data.

[0092] 403. Determine cover template features corresponding to the advertisement cover template data.

[0093] 404. Determine prompt words corresponding to each content element in the target advertisement cover according to the user preference characteristics and the cover template characteristics.

[0094] When generating prompt words corresponding to each content element, the user feedback data of the client on the historical advertisements that have been displayed can be obtained first. And the user preference features corresponding to the client can be determined based on the user portrait data and the user feedback data. The user preference features corresponding to the client can be determined by using a pre-trained user preference model.

[0095] Next, the cover template features corresponding to the advertisement cover template data are determined, wherein a pre-trained cover template model can be used to determine the cover template features corresponding to the advertisement cover template data.

[0096] Then, based on the user preference features and cover template features, the prompt words corresponding to each content element in the target advertisement cover are determined. Among them, as for the specific implementation method of determining the prompt words corresponding to each content element, in an optional embodiment, the current delivery scene and advertisement theme corresponding to the advertisement to be delivered can be first obtained; the current delivery scene, advertisement theme, user preference features and cover template features are input into the prompt word generation model to obtain the prompt words corresponding to each content element in the target advertisement cover.

[0097] In another optional embodiment, a prompt word template corresponding to the advertisement to be placed can be determined; description information corresponding to each content element in the target advertisement cover can be generated based on the current delivery scenario, advertisement theme, user preference characteristics and cover template characteristics; prompt words corresponding to each content element in the target advertisement cover can be generated based on the prompt word template and the description information corresponding to each content element.

[0098] The embodiment of the present invention determines the prompt words corresponding to each content element in the target advertisement cover according to the user preference characteristics and the cover template characteristics, thereby ensuring that the generated prompt words corresponding to each content element have both the user's individual characteristics and retain the main information in the original advertisement cover template.

[0099] In addition, in actual applications, in order to ensure that the generated target advertisement cover not only meets the personalized needs of users, but also complies with the advertisement specifications and the requirements set by advertisers, after generating the prompt words corresponding to each content element, the set constraints can also be used to optimize the prompt words.

[0100] In specific implementation, the target constraint conditions can be determined based on the user portrait data and the ad cover generation rules, and the prompt words corresponding to each content element can be optimized based on the target constraint conditions to obtain the optimized prompt words corresponding to each content element. Among them, the target constraint conditions include content compliance constraints, design aesthetic constraints, target audience constraints, and brand image constraints.

[0101] Among them, content compliance constraints can include laws and regulations, platform rules, ethical standards, etc., so as to ensure that the advertising content complies with local laws and regulations, such as advertising laws, copyright laws, etc. In addition, the platform rules for publishing advertisements, such as social media, search engines, etc., are also followed. At the same time, it can also ensure that the generated advertisement cover does not contain discriminatory, misleading or unethical information.

[0102] Among them, design aesthetic constraints can include constraints such as visual balance, color matching, and font selection, which can ensure that the layout of the visual elements of the advertising cover is balanced and harmonious, and can also limit the use of specific color schemes to maintain brand consistency or comply with industry practices. In addition, the type and size of the fonts used can also be restricted to improve readability and brand recognition.

[0103] Among them, the target audience constraints may include audience preferences, audience characteristics, audience behavior and other constraints, so that design elements and copywriting styles can be selected according to the preferences of the target audience, as well as combining factors such as the age, gender, cultural background, and purchasing power of the target audience. At the same time, the degree of personalization of the ad cover content can also be adjusted based on the audience's behavioral data.

[0104] Among them, brand image constraints can include constraints such as brand style, brand tone, and brand safety, so that the advertising design can be kept consistent with the brand visual identity system (VI), and the generated advertising copy and generated visual elements can be ensured to be consistent with the brand's market positioning and tone. At the same time, it can also avoid using content that may damage the brand image or content that may damage partners.

[0105] From the above description, it can be seen that: by using preset target constraints to optimize the prompt words corresponding to each generated content element, and then based on the prompt words corresponding to each optimized content element, guiding the advertising cover generation model to generate the target content corresponding to each content element, the generated target content can not only meet user preferences, but also comply with constraints such as advertising specifications and design aesthetics.

[0106] Among them, the specific implementation process involved in the embodiments of the present invention can refer to the contents in the above embodiments, and will not be repeated here.

[0107] In order to facilitate the understanding of the above embodiments, each of the above embodiments is described with examples in combination with specific application scenarios. Figure 5 As shown, when the client is detected to start the target application, the basic user information, user behavior data, and user feedback data corresponding to the client are collected. And based on the basic user information, user feedback data, and user behavior data corresponding to the client, the user portrait is analyzed to determine the user portrait data corresponding to the client. For example, the constructed user portrait data corresponding to client A is: User ID: ABCD; Gender: Male; Age: 28; Region: Beijing; Hobbies: Fashion, White, V-neck T-shirts; Purchase History: White V-neck T-shirts, Blue Jeans; Search History: White T-shirts, Summer Clothing.

[0108] Next, obtain the advertisement cover template corresponding to the advertisement to be placed, parse the advertisement cover template to extract the original key fields and multiple original design elements in the advertisement cover template, and generate advertisement cover template data based on the extracted original key fields and multiple original design elements. For example, the original key fields include copy, product pictures, buttons, etc. Among them, copy is a classic choice, which is versatile and elegant at every moment. The original design elements include background color and text color.

[0109] Then, based on the user portrait data and the advertisement cover template data, the prompt words corresponding to each content element in the target advertisement cover to be generated are determined. In addition, the prompt words corresponding to each content element in the target advertisement cover to be generated can also be determined in combination with the user feedback data.

[0110] Specifically, first, analyze user preferences based on user portrait data and user feedback data. And determine description information corresponding to multiple content elements that match user preferences based on collaborative filtering or content-based recommendations. Or use a rule engine to match description information corresponding to multiple content elements based on preset rules. Based on the description information corresponding to the multiple content elements, generate prompt words corresponding to the multiple content elements. For example, generate background prompt words, picture prompt words, and text prompt words.

[0111] Next, multiple design elements corresponding to the target advertisement cover are determined based on the advertisement cover template data and the advertisement cover generation rules corresponding to the advertisement to be placed. Display information corresponding to the multiple design elements is determined based on the user portrait data. Prompt words corresponding to the multiple design elements are generated based on the display information corresponding to the multiple design elements.

[0112] At the same time, in order to ensure that the generated target advertising cover is consistent with the advertising theme and current delivery scenario of the advertisement to be delivered, scene prompt words, advertising theme prompt words, user portrait prompt words, advertising template prompt words, generation requirement prompt words corresponding to design elements, generation requirement prompt words corresponding to content elements, etc. can also be generated at the same time.

[0113] Next, target constraints are determined based on user portrait data and ad cover generation rules. Based on the target constraints, the prompt words corresponding to each content element and the prompt words corresponding to multiple design elements are optimized to obtain optimized prompt words corresponding to each content element and optimized prompt words corresponding to each design element, so as to ensure that the generated target content meets the advertising goals and user preferences.

[0114] The optimized prompt words corresponding to each content element and the optimized prompt words corresponding to multiple design elements, scene prompt words, advertising theme prompt words, user portrait prompt words, advertising template prompt words, generation requirement prompt words corresponding to design elements and generation requirement prompt words corresponding to content elements are merged to integrate multiple prompt words into one target prompt word.

[0115] Finally, the target prompt words are input into the advertisement cover generation model to generate the target content corresponding to each content element. According to the target content corresponding to each content element, the target advertisement cover corresponding to the client is generated to display the target advertisement cover in the target application corresponding to the client.

[0116] Among them, when combining various content elements to generate a target advertising cover, a dynamic combination optimization algorithm can be used to layout multiple content elements to determine the optimal layout method, and multiple target contents can be combined according to the optimal layout method, so that the generated target advertising cover can be more beautiful.

[0117] Among them, genetic algorithms can be used to optimize the combination of multiple content elements, and the optimal solution can be found by simulating natural selection and genetic mechanisms. Or through reinforcement learning, the optimal strategy can be learned by interacting with the environment to optimize the combination of multiple content elements. Or simulated annealing can be used to find the global optimal solution under the set constraints to determine the optimal typesetting method.

[0118] From the above description, it can be seen that: by generating prompt words corresponding to each content element in the advertisement cover based on user portrait data and the advertisement cover template data corresponding to the advertisement to be placed, and integrating the prompt words corresponding to each content element, the advertisement cover generation model can be better guided to accurately control the generation of each content element according to the prompt words corresponding to each content element, so that each content element in the generated target advertisement cover can be accurately presented according to the design effect corresponding to the prompt words. This can not only ensure that the generated target advertisement cover is highly relevant to the user's interests and needs, so as to customize the personalized advertisement cover corresponding to the client, but also ensure the quality of the generated target advertisement cover.

[0119] The following will describe in detail the advertisement cover generation device of one or more embodiments of the present invention. Those skilled in the art will appreciate that the advertisement cover generation device can be configured using commercially available hardware components through the steps taught in this solution.

[0120] Figure 6 A schematic diagram of the structure of an advertisement delivery device provided by an embodiment of the present invention is shown in FIG. Figure 7 As shown, the device includes: an acquisition module 11, a determination module 12, a fusion module 13, a first generation module 14, and a second generation module 15.

[0121] The acquisition module 11 is used to acquire the user portrait data corresponding to the client and the advertisement cover template data corresponding to the advertisement to be placed if it is detected that the client starts the target application.

[0122] The determination module 12 is used to determine the prompt words corresponding to each content element in the target advertisement cover to be generated according to the user portrait data and the advertisement cover template data.

[0123] The fusion module 13 is used to fuse the prompt words corresponding to the various content elements, so as to integrate the prompt words corresponding to the content elements into a target prompt word.

[0124] The first generating module 14 is used to input the target prompt word into the advertisement cover generating model to generate the target content corresponding to each content element.

[0125] The second generating module 15 is used to generate a target advertisement cover corresponding to the client according to the target content corresponding to each content element, so as to display the target advertisement cover in a target application corresponding to the client.

[0126] In an optional embodiment, the fusion module 13 is specifically used to: determine multiple design elements corresponding to the target advertisement cover according to the advertisement cover template data and the advertisement cover generation rules corresponding to the advertisement to be placed; determine display information corresponding to the multiple design elements according to the user portrait data; generate prompt words corresponding to the multiple design elements according to the display information corresponding to the multiple design elements; and fuse the prompt words corresponding to the respective content elements and the prompt words corresponding to the multiple design elements to integrate the prompt words corresponding to the respective content elements and the prompt words corresponding to the multiple design elements into one target prompt word.

[0127] In an optional embodiment, the fusion module 13 is specifically used to: optimize the display information corresponding to the multiple design elements by simulating natural selection and genetic mechanisms to obtain the optimized target display information corresponding to the multiple design elements; and generate prompt words corresponding to the multiple design elements according to the target display information corresponding to the multiple design elements.

[0128] In an optional embodiment, before the prompt words corresponding to the various content elements are merged to integrate the prompt words corresponding to the various content elements into a target prompt word, the fusion module 13 is also used to: determine target constraint conditions according to the user portrait data and the advertisement cover generation rules; optimize the prompt words corresponding to the various content elements according to the target constraint conditions to obtain optimized prompt words corresponding to the various content elements.

[0129] In an optional embodiment, the target constraint conditions include content compliance constraints, design aesthetic constraints, target audience constraints, and brand image constraints.

[0130] In an optional embodiment, the determination module 12 is specifically used to: obtain user feedback data of the client on historical advertisements that have been displayed; determine user preference characteristics corresponding to the client based on the user portrait data and the user feedback data; determine cover template characteristics corresponding to the advertisement cover template data; and determine prompt words corresponding to each content element in the target advertisement cover based on the user preference characteristics and the cover template characteristics.

[0131] In an optional embodiment, the determination module 12 is specifically used to: obtain the current delivery scenario and advertising theme corresponding to the advertisement to be delivered; input the current delivery scenario, the advertising theme, the user preference characteristics and the cover template characteristics into a prompt word generation model to obtain the prompt words corresponding to each content element in the target advertisement cover.

[0132] In an optional embodiment, the determination module 12 is specifically used to: determine a prompt word template corresponding to the advertisement to be placed; generate description information corresponding to each content element in the target advertisement cover according to the current delivery scenario, the advertisement theme, the user preference characteristics and the cover template characteristics; generate prompt words corresponding to each content element in the target advertisement cover according to the prompt word template and the description information corresponding to each content element.

[0133] Figure 6 The device shown can execute the steps in the advertising cover generation method in the aforementioned embodiment. The detailed execution process and technical effects can be found in the description in the aforementioned embodiment and will not be repeated here.

[0134] The embodiment of the present invention also provides an electronic device, such as Figure 7 As shown, the electronic device may include: a processor 21, a memory 22, and a communication interface 23. The memory 22 stores executable codes, and when the executable codes are executed by the processor 21, the processor 21 implements the advertisement cover generation method in the above-mentioned embodiment.

[0135] In addition, an embodiment of the present invention provides a non-temporary machine-readable storage medium, on which executable code is stored. When the executable code is executed by a processor of an electronic device, the processor can at least implement the advertising cover generation method provided in the aforementioned embodiment.

[0136] In addition, an embodiment of the present invention provides a computer program product having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor can at least implement the advertisement cover generation method provided in the aforementioned embodiment.

[0137] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. Those of ordinary skill in the art may understand and implement the present invention without creative effort.

[0138] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by adding a necessary general hardware platform, and of course can also be implemented by combining hardware and software. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a computer product, and the present invention can be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating an advertisement cover, characterized in that: include: If it is detected that the client starts the target application, the user portrait data corresponding to the client and the advertisement cover template data corresponding to the advertisement to be placed are obtained; Determine, according to the user portrait data and the advertisement cover template data, prompt words corresponding to each content element in the target advertisement cover to be generated; Merging the prompt words corresponding to the various content elements to integrate the prompt words corresponding to the various content elements into a target prompt word; Inputting the target prompt words into the advertisement cover generation model to generate target content corresponding to each content element; According to the target content corresponding to each content element, a target advertisement cover corresponding to the client is generated, so as to display the target advertisement cover in a target application corresponding to the client.

2. The method according to claim 1, characterized in that The step of fusing the prompt words corresponding to the various content elements to integrate the prompt words corresponding to the various content elements into a target prompt word includes: Determine a plurality of design elements corresponding to the target advertisement cover according to the advertisement cover template data and the advertisement cover generation rule corresponding to the advertisement to be placed; Determining display information corresponding to the multiple design elements according to the user portrait data; Generating prompt words corresponding to the multiple design elements according to the display information corresponding to the multiple design elements; The prompt words corresponding to the respective content elements and the prompt words corresponding to the plurality of design elements are merged to integrate the prompt words corresponding to the respective content elements and the prompt words corresponding to the plurality of design elements into a target prompt word.

3. The method according to claim 2, characterized in that The step of generating prompt words corresponding to the plurality of design elements according to the display information corresponding to the plurality of design elements includes: By simulating natural selection and genetic mechanisms, the display information corresponding to the multiple design elements is optimized to obtain optimized target display information corresponding to the multiple design elements; According to the target display information corresponding to the multiple design elements, prompt words corresponding to the multiple design elements are generated.

4. The method according to claim 1, characterized in that Before fusing the prompt words corresponding to the various content elements to integrate the prompt words corresponding to the various content elements into a target prompt word, the method further includes: Determining target constraint conditions according to the user portrait data and the advertisement cover generation rules; According to the target constraint condition, the prompt words corresponding to the various content elements are optimized to obtain the optimized prompt words corresponding to the various content elements.

5. The method according to claim 4, characterized in that The target constraints include content compliance constraints, design aesthetic constraints, target audience constraints, and brand image constraints.

6. The method according to claim 1, characterized in that The step of determining, based on the user portrait data and the advertisement cover template data, prompt words corresponding to the content elements in the target advertisement cover to be generated includes: Obtaining user feedback data of the client on historical advertisements that have been displayed; Determining user preference features corresponding to the client according to the user portrait data and the user feedback data; Determine the cover template features corresponding to the advertisement cover template data; According to the user preference feature and the cover template feature, the prompt words corresponding to the various content elements in the target advertisement cover are determined.

7. The method according to claim 6, characterized in that The step of determining the prompt words corresponding to the various content elements in the target advertisement cover according to the user preference feature and the cover template feature includes: Obtaining the current delivery scenario and advertisement theme corresponding to the advertisement to be delivered; The current delivery scenario, the advertisement theme, the user preference features and the cover template features are input into a prompt word generation model to obtain prompt words corresponding to each content element in the target advertisement cover.

8. The method according to claim 6, characterized in that The step of determining the prompt words corresponding to the various content elements in the target advertisement cover according to the user preference feature and the cover template feature includes: Determine a prompt word template corresponding to the advertisement to be placed; Generate description information corresponding to each content element in the target advertisement cover according to the current delivery scenario, the advertisement theme, the user preference characteristics and the cover template characteristics; The prompt words corresponding to the respective content elements in the target advertisement cover are generated according to the prompt word template and the description information corresponding to the respective content elements.

9. An electronic device, characterized in that: include: A memory, a processor, and a communication interface; wherein the memory stores executable code, and when the executable code is executed by the processor, the processor executes the advertising cover generation method as described in any one of claims 1 to 8.

10. A non-transitory machine-readable storage medium, characterized in that: The non-transitory machine-readable storage medium stores executable code, and when the executable code is executed by a processor of an electronic device, the processor executes the advertisement cover generation method as described in any one of claims 1 to 8.

Images