Intelligent poster generation method, device and equipment based on large model

Abstract

The embodiment of the invention provides an intelligent poster generation method, device and equipment based on a large model, and the method comprises the steps: obtaining the initial template data of each poster template stored in a target platform database, carrying out the text processing of the initial template data, determining the annotation information corresponding to each target text data, and obtaining the annotation template data; classifying the annotation template data according to annotation information of the annotation template data, and establishing a structured data table; in response to the received poster creation data submitted by the target object, constructing a retrieval condition; and determining target template data according to the retrieval condition and the structured data table, and generating a target poster. And the efficiency of generating the intelligent poster is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer technology and related technical fields, in particular, to a large model-based intelligent poster generation method, device and equipment. BACKGROUND

[0002] With the development of Internet technology, people can not only browse various information through the network, but also share information in various ways. For example, in the field of financial technology, posters can be used to share information such as enterprise business, achievements of enterprises, cultural image of enterprises, and activity organization propaganda within enterprises.

[0003] The existing poster generation method has the following problems: (1) slow poster generation speed, traditional poster design usually needs professional designers to spend a lot of time for design and modification, and cannot quickly respond to market changes and emergencies; (2) high design threshold, professional design software and design knowledge have a certain learning threshold for front-line personnel, and they cannot independently complete poster design without professional design skills; (3) hot spot response lag, hot events have timeliness, if the poster cannot be published in time, the opportunity brought by the hot spot cannot be grasped; (4) difficulty in personalized customization, traditional batch poster generation method cannot meet the personalized needs of different users; (5) lack of intelligence, the existing technology lacks intelligent assistance in the poster design process, such as automatic layout optimization and color matching, which affects the quality and aesthetics of the poster.

[0004] Based on the problems existing in the prior art, there is an urgent need for a solution that can quickly, simply, intelligently and personalized generate posters, so as to solve the problem that front-line personnel cannot quickly generate hot spot posters. SUMMARY

[0005] The embodiments described herein provide a large model-based intelligent poster generation method, device, equipment and medium, which solves the problems existing in the prior art.

[0006] In a first aspect, according to the content of the present disclosure, a large model-based intelligent poster generation method is provided, comprising: obtaining initial template data of each poster template stored in a target platform database, performing text processing on the initial template data, and determining annotation information corresponding to each target text data to obtain annotated template data; classifying the annotated template data according to the annotation information of the annotated template data, and establishing a structured data table, wherein the structured data table includes template identification, annotation information and target text data; In response to receiving poster creation data submitted by a target object, constructing a retrieval condition; According to the search condition and the structured data table, target template data is determined, and a target poster is generated.

[0007] In some embodiments of the present disclosure, the initial template data of each poster template stored in the target platform database is acquired, text processing is performed on the initial template data, and the label information corresponding to each target text data is determined to obtain labeled template data, including: The initial template data of each poster template stored in the target platform database is acquired, and text data in the initial template data is extracted; According to the position information of the text data in the initial template data, the text data is sorted; According to the semantic information of each sorted text data, the correlation of each text data is determined, and the text data with correlation is textually associated to obtain target text data; According to the semantic information of each target text data, the label information of each target text data is determined; According to the label information of each target text data and each target text data, labeled template data is obtained.

[0008] In some embodiments of the present disclosure, the labeled template data is classified according to the label information of the labeled template data, and a structured data table is established, including: The labeled template data is classified according to the template classification label information of the labeled template data to obtain labeled template data of different template classifications; According to the labeled template data included in different template classifications, a structured data table of template identification, template architecture label information and target text data is constructed.

[0009] In some embodiments of the present disclosure, the labeled template data of different template classifications is obtained by classifying the labeled template data according to the template classification label information of the labeled template data, including: According to the template classification label information of the labeled template data, the first-level label and the second-level label of the template classification corresponding to each labeled template data are determined; According to the first-level label and the second-level label of the template classification corresponding to each labeled template data, the labeled template data is classified to obtain labeled template data of different template classifications.

[0010] In some embodiments of the present disclosure, in response to receiving poster creation data submitted by a target object, a search condition is constructed according to the poster creation data, including: In response to receiving poster creation data submitted by a target object, the poster creation data is processed to obtain a plurality of poster creation word groups; According to the poster creation word group, a target search keyword and target annotation information corresponding to the target search keyword are determined; A search condition is constructed according to the target search keyword and the target annotation information corresponding to the target search keyword.

[0011] In some embodiments of the present disclosure, the search condition is constructed according to the target search keyword and the target annotation information corresponding to the target search keyword, comprising: According to the target annotation information corresponding to the target search keyword, the target annotation information corresponding to the target search keyword is classified to obtain target template classification annotation information and target template architecture annotation information; Based on the target template classification annotation information and the target search keyword corresponding to the target template classification annotation information, a first search condition is constructed; Based on the target template architecture annotation information and the target search keyword corresponding to the target template architecture annotation information, a second search condition is constructed based on the first search condition.

[0012] In some embodiments of the present disclosure, the target template data is determined and the target poster is generated according to the search condition and the structured data table, comprising: According to the target template classification annotation information included in the search condition and the target search keyword corresponding to the target template classification annotation information, the target template data is obtained from the structured data table; According to the target template architecture annotation information included in the search condition and the target search keyword corresponding to the target template architecture annotation information, target text data corresponding to the target template architecture annotation information is screened from the target template data; According to the matching degree of the target text data corresponding to the target template architecture annotation information and the target search keyword corresponding to the target template architecture annotation information, initial target template data is determined; According to the initial target template data and the target search keyword corresponding to the target template architecture annotation information, target template data is determined, and a target poster is generated based on the target template data.

[0013] In some embodiments of the present disclosure, the target template data is determined according to the initial target template data and the target search keyword corresponding to the target template architecture annotation information, comprising: Hot information is obtained and analyzed to determine attribute information corresponding to each hot information; According to the attribute information corresponding to each hot information, the relevance of each hot information to the poster creation data submitted by the target object is determined; According to the relevance of each hot information to the poster creation data submitted by the target object, target hot information satisfying a preset relevance is selected; According to the initial target template data, the target search keyword corresponding to the target template architecture annotation information, and the target hot spot information, determine target template data.

[0014] In a second aspect, according to the content of the disclosure, an intelligent poster generation device based on a large model is provided, comprising: A data processing module is configured to obtain initial template data of each poster template stored in a target platform database, perform text processing on the initial template data, and determine annotation information corresponding to each target text data to obtain annotated template data. A data table construction module is configured to classify the annotated template data according to the annotation information of the annotated template data, and construct a structured data table, wherein the structured data table includes template identification, annotation information, and target text data. A search condition construction module is configured to construct a search condition in response to receiving poster creation data submitted by a target object. A poster generation module is configured to determine target template data according to the search condition and the structured data table, and generate a target poster.

[0015] In a third aspect, according to the content of the disclosure, a computer device is provided, comprising: One or more processors; A storage device configured to store one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of the first aspect.

[0016] The method, device and equipment for generating an intelligent poster based on a large model provided by the embodiments of the present disclosure first acquire initial template data of each poster template stored in a target platform database, perform text processing on the initial template data, and determine label information corresponding to each target text data to obtain labeled template data; then, according to the label information of the labeled template data, the labeled template data is classified, and a structured data table is established; further, in response to receiving poster creation data submitted by a target object, a retrieval condition is constructed; finally, according to the retrieval condition and the structured data table, target template data is determined, and a target poster is generated. On the one hand, by acquiring the initial template data of each poster template provided by the target platform, and constructing the structured data table of each initial template data, the efficiency of subsequent target poster template retrieval is facilitated; on the other hand, after receiving the poster creation data, based on the template classification requirements of the target object, a plurality of target labeled template data meeting the template classification requirements of the target object are filtered out from the structured data table, and then the target retrieval keywords corresponding to the main title or text template label information provided by the target object are matched with the plurality of target labeled template data, so as to realize accurate screening of the initial target template data, finally, the target retrieval keywords corresponding to the label information are replaced with the target text content corresponding to the label information in the initial target template data, to obtain the target template data, and an intelligent poster is generated.

[0017] The above description is only a summary of the technical solutions of the embodiments of the present application. In order to more clearly understand the technical means of the embodiments of the present application, the embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the embodiments of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly described below. It should be noted that the drawings described below only relate to some embodiments of the present disclosure, but not limit the present disclosure, wherein: Figure 1 is a flow diagram of a method for generating an intelligent poster based on a large model provided by an embodiment of the present disclosure; Figure 2 is a structural diagram of an intelligent poster generation device based on a large model provided by an embodiment of the present disclosure; Figure 3 is a structural diagram of a computer device provided by an embodiment of the present disclosure.

[0019] In the drawings, the last two digits of the same reference signs correspond to the same elements. It should be noted that the elements in the drawings are schematic and not drawn to scale. DETAILED DESCRIPTION

[0020] In order to make the purposes, technical solutions and advantages of the embodiments of the present disclosure clearer, the technical solutions of the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present disclosure.

[0021] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this present subject matter belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the specification and relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein. As used herein, the statement that two or more parts are "connected" or "coupled" together refer to an indirect or direct connection or coupling.

[0022] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase "in an embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily all directed to the same embodiment, or to a single alternative embodiment. It is explicitly contemplated that embodiments described herein can be combined with each other.

[0023] The term "and / or", merely an associative relationship of the associated objects, means that there can be three relationships, for example, A and / or B, which can represent: there is A, there is A and B, and there is B. In addition, the character " / " herein generally represents that the front and rear associated objects are a "or" relationship.

[0024] In addition, in all embodiments of the present disclosure, terms such as "first" and "second" are only used to distinguish one component (or part of the component) from another component (or another part of the component).

[0025] In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more (including two), and similarly, "a plurality of groups" means two or more groups (including two groups).

[0026] In order to make the persons skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings.

[0027] Based on the problems existing in the prior art, the embodiment of the present disclosure provides a large model-based intelligent poster generation method, Figure 1 is a flowchart of a large model-based intelligent poster generation method provided by the embodiment of the present disclosure, as Figure 1 shown, the large model-based intelligent poster generation method comprises: S110, obtaining initial template data of each poster template stored in a target platform database, performing text processing on the initial template data, and determining annotation information corresponding to each target text data to obtain annotated template data.

[0028] Among them, the target platform is a private domain platform, and the database of the private domain platform stores initial template data of each poster template.

[0029] In the large model-based intelligent poster generation method provided by the embodiment of the present disclosure, first, the initial template data of each poster template stored in the target platform database is obtained, then the text data extraction is performed on the initial template data based on the large model, and the semantic recognition is performed on the extracted text data, the annotation information corresponding to each text data is determined, and the annotated template data is generated.

[0030] In a specific implementation manner, obtaining initial template data of each poster template stored in a target platform database, performing text processing on the initial template data, and determining annotation information corresponding to each target text data to obtain annotated template data comprises: obtaining initial template data of each poster template stored in a target platform database, and extracting text data in the initial template data; according to the position information of the text data in the initial template data, the text data is sorted; according to the semantic information of each text data after sorting, the correlation relationship of each text data is determined, and the text data with the correlation relationship is text associated to obtain target text data; according to the semantic information of each target text data, the annotation information of each target text data is determined; according to the annotation information of each target text data and each target text data, the annotated template data is obtained.

[0031] Before initiating the demand for generating a poster, the target object needs to complete the intelligent annotation of the initial template data first. The core target is to clarify the semantic information of the text data of each poster template, and based on the determined semantic information of the text data of each poster template, the annotation information of each text data is determined. The specific implementation manner is as follows: The initial template data of each poster template stored in the database of the target platform is first extracted all the text data included in each poster template, and then sorted according to the coordinate values of the text data, and the sorting rule follows from top to bottom and from left to right, avoiding semantic understanding errors of large models due to chaotic text data order, and improving the accuracy of the recognition results of the associated semantic text data. After sorting the extracted text data, the sorted text data is subjected to semantic recognition based on the large model, and based on the semantic recognition results, the text data with associated relationship is subjected to text association to obtain target text data. In addition, in the process of performing semantic recognition on each text data based on the large model and associating the text data with associated relationship to obtain target text data, the annotation information of each target text data is determined according to the semantic information of each target text data, and finally the annotation information of the target text data corresponding to each poster template and the target text data form the annotated template data.

[0032] A specific example, the large model automatically identifies the semantic association relationship between each text data, and completes the annotation of the associated text data. For example, "seven" and "Tian speed class" are associated texts, then "seven" and "Tian speed class" form a target text data "seven-day speed class", based on the semantic understanding of the target text data "seven-day speed class", the annotation information of the target text data is determined as "course duration description".

[0033] It should be noted that the initial template data of each poster template only includes the text data included in each poster template and the position information corresponding to each text data, and the annotated template data of each poster template not only includes the text data included in each poster template and the position information corresponding to each text data, but also includes the annotation information corresponding to each text data.

[0034] In addition, it should be further pointed out that the annotation information includes not only the main title, the text, the time and other template architecture annotation information, but also the template classification annotation information. The template classification includes template type, template channel, template purpose, template industry, template color, template style, template scene, etc.

[0035] S120, according to the annotation information of the annotated template data, the annotated template data is classified, and a structured data table is established.

[0036] Among them, the structured data table includes template identification, annotation information and target text data; It should be noted that the template classification includes template type, template channel, template purpose, template industry, template color, template style, template scene, etc., wherein the template type includes poster, long picture poster, short video poster, the template channel includes e-commerce, short video, etc., the template purpose includes happy report, battle report, etc., the template industry includes food industry, retail industry, clothing industry, etc., the template color includes Mid-Autumn theme color, Spring Festival main color, etc., the template style includes national trend style, simple style, etc., and the template scene includes promotion scene, sprint scene, etc.

[0037] The same poster template can correspond to multiple template classifications, that is, the template classification of the poster template 1 can be understood as: the template type is a long picture poster, the template purpose is a happy report, the template industry is a food industry,..., and the template classification of the poster template n is: the template type is a short video poster, the template channel is e-commerce, and the template color is a Mid-Autumn theme color.

[0038] To ensure the efficiency of subsequent retrieval of the target poster template, in the intelligent poster generation method based on a large model provided in the embodiments of the present disclosure, before generating the target poster template, the labeled template data is classified according to the labeling information of the labeled template data, and a structured data table corresponding to different template classifications is established.

[0039] In specific embodiments, the labeled template data is classified according to the labeling information of the labeled template data, and a structured data table is established, including: classifying the labeled template data according to the template classification labeling information of the labeled template data to obtain labeled template data of different template classifications; and constructing a structured data table of template identification, template architecture labeling information and target text data according to the labeled template data included in different template classifications.

[0040] According to the template classification labeling information of the labeled template data, the labeled template data is classified to obtain labeled template data of different template classifications, including: determining the first-level label and the second-level label of the template classification corresponding to each labeled template data according to the template classification labeling information of the labeled template data; and classifying the labeled template data according to the first-level label and the second-level label of the template classification corresponding to each labeled template data to obtain labeled template data of different template classifications.

[0041] Since the template classification includes template type, template channel, template purpose, template industry, template color, template style, template scene, and the template type, template channel, template purpose, template industry, template color, template style, template scene include multiple different sub-classifications, the structured data table constructed can be understood as the template type, template channel, template purpose, template industry, template color, template style, template scene as multiple parallel first-level labels, and the multiple different sub-classifications included in the template type, template channel, template purpose, template industry, template color, template style, template scene as second-level labels under the corresponding first-level labels. It can be understood that the template type, template channel, template purpose, template industry, template color, template style, template scene are multiple parallel first-level labels under the template classification, and the poster, long-poster, short-video poster are three parallel second-level labels under the first-level label template type, the e-commerce, short video are two parallel second-level labels under the first-level label template channel, and the promotion scene, sprint e-commerce scene are two parallel second-level labels under the first-level label template scene.

[0042] If the annotation information of the target text data of a certain poster template determined in step S110 includes the promotion scene annotation information, the annotation template data of the poster template is stored in the template classification corresponding to the template scene-promotion scene. If the annotation information of the target text data of a certain poster template determined includes both the promotion scene annotation information and the food industry annotation information, the annotation template data of the poster template is stored in the template classification corresponding to the template scene-promotion scene and also in the template industry-food industry classification.

[0043] It should be noted that in the above implementation process, there may be a template classification annotation information of the target text data of a certain poster template corresponding to the first-level label of the template classification, and then the annotation template data of the poster template can be stored under the corresponding first-level label of the template classification.

[0044] The data of the poster templates stored in the same template classification are assigned template identifiers in the order of storage.

[0045] Based on the above classification of the annotation template data of each poster template according to the template classification, a structured data table of template identifier, template architecture annotation information, and target text data is constructed according to the annotation template data included in different template classifications.

[0046] In addition, after constructing the structured data table of each template classification, each data in the structured data table is converted into a vector (using vector embedding technology to convert text data into a vector format recognizable by a computer), and finally stored in a database to provide efficient data support for subsequent template retrieval.

[0047] S130, in response to receiving the poster creation data submitted by the target object, constructing a retrieval condition.

[0048] In a specific embodiment, in response to receiving the poster creation data submitted by the target object, constructing a retrieval condition according to the poster creation data includes: in response to receiving the poster creation data submitted by the target object, performing word segmentation on the poster creation data to obtain a plurality of poster creation word groups; determining target retrieval keywords and target annotation information corresponding to the target retrieval keywords according to the poster creation word groups; and constructing a retrieval condition according to the target retrieval keywords and the target annotation information corresponding to the target retrieval keywords.

[0049] Based on the poster creation data submitted by the target object, the retrieval intention is accurately identified through the large model processing, and then the retrieval condition is constructed. The specific process is as follows: After receiving the poster creation data submitted by the target object, the poster creation data is processed to obtain a plurality of poster creation word groups. The semantic recognition of the large model is performed on the plurality of poster creation word groups, the target retrieval keywords are selected from the plurality of poster creation word groups, and the target annotation information corresponding to the target retrieval keyword group is determined based on the semantic information of the target retrieval keyword.

[0050] A specific example is that the poster creation data input by the target object is "I want to make a double eleven e-commerce poster". After word segmentation of the poster creation data, a plurality of poster creation word groups are obtained, which are "I", "want to", "make a", "double eleven", "e-commerce", and "poster". Through semantic recognition of each poster creation word group, target retrieval keywords related to the target poster template are selected, for example, the target retrieval keywords determined from the plurality of poster creation word groups include "double eleven", "e-commerce", and "poster". Then, based on the semantic information of each target retrieval keyword, the target annotation information of "double eleven" is determined as a keyword, the target annotation information of "e-commerce" is determined as a template channel, and the target annotation information of "poster" is determined as a template type.

[0051] Among them, constructing a retrieval condition according to the target retrieval keywords and the target annotation information corresponding to the target retrieval keywords includes: classifying the target annotation information corresponding to the target retrieval keywords according to the target annotation information corresponding to the target retrieval keywords to obtain target template classification annotation information and target template architecture annotation information; based on the target template classification annotation information and the target retrieval keywords corresponding to the target template classification annotation information, constructing a first retrieval condition; based on the target template architecture annotation information and the target retrieval keywords corresponding to the target template architecture annotation information, constructing a second retrieval condition based on the first retrieval condition.

[0052] After the target search keyword and the target annotation information corresponding to the target search keyword are determined, a search condition is constructed according to the target search keyword and the target annotation information corresponding to the target search keyword. In combination with the above embodiment, it can be determined that the annotation information is a template channel and the template type is target template classification annotation information, and the annotation information is a keyword, which is target template architecture annotation information. Then, based on the target template classification annotation information and the target search keyword corresponding to the target template classification annotation information, the first search condition constructed is {template classification-template channel-e-commerce} and {template classification-template type-poster}. Based on the first search condition, based on the target template architecture annotation information and the target search keyword corresponding to the target template architecture annotation information, the second search condition constructed is {template classification-template channel-e-commerce, keyword: Double 11} and {template classification-template type-poster, keyword: Double 11}.

[0053] In a specific implementation manner, the target template data is determined according to the search condition and the structured data table, and the target poster is generated.

[0054] In a specific implementation manner, the target template data is determined according to the search condition and the structured data table, and the target poster is generated. Specifically, the target annotation template data is obtained from the structured data table according to the target template classification annotation information included in the search condition and the target search keyword corresponding to the target template classification annotation information. The target text data corresponding to the target template architecture annotation information is selected from the target annotation template data according to the target template architecture annotation information included in the search condition and the target search keyword corresponding to the target template architecture annotation information. The initial target template data is determined according to the matching degree between the target text data corresponding to the target template architecture annotation information and the target search keyword corresponding to the target template architecture annotation information. The target template data is determined according to the initial target template data and the target search keyword corresponding to the target template architecture annotation information, and the target poster is generated based on the target template data. After determining the target retrieval keyword and the target annotation information corresponding to the target retrieval keyword, first, the target template classification annotation information and the target retrieval keyword corresponding to the target template classification annotation information are used to filter target annotation template data from the structured data table. In the above embodiment, for example, the first-level label of the filtered target annotation template data is a template channel, and the second-level label is e-commerce, and the first-level label of the filtered target annotation template data is a template type, and the second-level label is a poster. Then, the target template architecture annotation information included in the retrieval condition and the target retrieval keyword corresponding to the target template architecture annotation information are used to filter target text data corresponding to the target template architecture annotation information from the target annotation template data. In the above embodiment, for example, the similarity between the target text data corresponding to the target template architecture annotation information and the "Double 11" text data is calculated, the initial target template data corresponding to the target text data with the highest similarity is selected, the "Double 11" target retrieval keyword is used to replace the target text data in the keyword annotation information of the initial target template data, and the target template data is generated. Then, the target poster is generated based on the target template data.

[0055] In the specific implementation process, by establishing a target platform template search interface parameter set, the retrieval condition of the large model after expansion is corresponded to the target platform template search interface parameter through a dictionary mapping relationship, a real user retrieval scene is simulated, and the template matching accuracy is improved.

[0056] The method provided in the embodiments of the present disclosure first acquires initial template data of each poster template stored in a target platform database, performs text processing on the initial template data, determines annotation information corresponding to each target text data, and obtains annotated template data. Then, according to the annotation information of the annotated template data, the annotated template data is classified, and a structured data table is established. Furthermore, in response to receiving poster creation data submitted by a target object, a retrieval condition is constructed. Finally, according to the retrieval condition and the structured data table, target template data is determined, and a target poster is generated. On the one hand, by acquiring initial template data of each poster template provided by a target platform and constructing a structured data table of each initial template data, the efficiency of subsequent target poster template retrieval is facilitated. On the other hand, after receiving poster creation data, based on the template classification requirements of the target object, a plurality of target annotation template data satisfying the template classification requirements of the target object are filtered from the structured data table. Then, the target retrieval keyword corresponding to the main title or text template annotation information provided by the target object is matched with the plurality of target annotation template data, the initial target template data is accurately filtered, and finally the target retrieval keyword corresponding to the annotation information is replaced with the target text content corresponding to the annotation information in the initial target template data, the target template data is obtained, and the poster is intelligently generated.

[0057] As a preferred implementation manner, on the basis of the above embodiment, before performing target template data determination according to the target search keyword corresponding to the initial target template data and the target template architecture annotation information, the intelligent poster generation method based on a large model provided by the embodiment of the disclosure further comprises: Obtaining hot spot information, and analyzing the hot spot information to determine attribute information corresponding to each hot spot information; determining the relevance of each hot spot information and the poster creation data submitted by the target object according to the attribute information corresponding to each hot spot information; selecting target hot spot information with relevance satisfying a preset relevance according to the relevance of each hot spot information and the poster creation data submitted by the target object; and determining the target template data according to the target search keyword corresponding to the initial target template data and the target template architecture annotation information and the target hot spot information.

[0058] To ensure that the generated target poster meets the real-time hot spot demand, the TrendRadar open source tool is used to realize automatic acquisition and update of hot spot data, and the specific principle and process are as follows: based on the technical scheme of “RSS subscription + LLM analysis”, the TrendRadar integrates the RSS subscription interface of multiple mainstream news sources, crawls the hot spot information on the Internet in real time, and identifies the hot spot attributes (such as “holiday hot spot”, “social event hot spot”, “industry hot spot”, etc.) of the crawled content through LLM. The identified hot spot information is pushed to the enterprise chat tool, and after the operator clicks “confirm”, the hot spot is automatically updated to the “hot spot pool” (hot spot database), providing real-time hot spot materials for subsequent poster copy generation.

[0059] Further, based on the poster creation data submitted by the target object and the hot spot data, a copy that meets the demand is generated, and the poster layout is optimized, which is implemented as follows: The historical annotation content (such as “promotion copy position annotation” and “title text annotation”) of the initial target template data is extracted, combined with the real-time hot spot (such as “Mid-Autumn Festival hot spot: reunion, discount”) in the hot spot pool, and a copy that meets the scene and user demand is generated through a large model (such as the title “Mid-Autumn Festival reunion season, food 300 yuan minus 100”). Further, a layout rule engine is introduced, which dynamically adjusts the template annotation according to the length and semantic attributes of the AI-generated copy (such as automatically deleting redundant decorative annotations when the copy is too long, and adding auxiliary information annotations when the copy is insufficient); at the same time, the engine automatically identifies the text background and decorative elements (such as borders and patterns) in the template to ensure that the generated poster meets the picture style of the Little Red Book platform (such as high saturation color, lightweight decoration, and concise and eye-catching copy), balancing aesthetics and platform adaptability.

[0060] On the basis of the above-mentioned embodiments, the disclosure embodiments further provide a large model-based intelligent poster generation device, Figure 2 is a structural schematic diagram of a large model-based intelligent poster generation device provided by the disclosure embodiments, as Figure 2 indicated, the large model-based intelligent poster generation device comprises: The data processing module 210 is configured to obtain initial template data of each poster template stored in a target platform database, perform text processing on the initial template data, and determine annotation information corresponding to each target text data to obtain annotated template data. The data table construction module 220 is configured to classify the annotated template data according to the annotation information of the annotated template data, and construct a structured data table, wherein the structured data table comprises template identification, annotation information and target text data. The retrieval condition construction module 230 is configured to construct a retrieval condition in response to receiving poster creation data submitted by a target object. The poster generation module 240 is configured to determine target template data according to the retrieval condition and the structured data table, and generate a target poster.

[0061] The large model-based intelligent poster generation device provided by the disclosure embodiments first obtains initial template data of each poster template stored in a target platform database, performs text processing on the initial template data, and determines annotation information corresponding to each target text data to obtain annotated template data. Then, according to the annotation information of the annotated template data, the annotated template data is classified, and a structured data table is constructed. Then, in response to receiving poster creation data submitted by a target object, a retrieval condition is constructed. Finally, according to the retrieval condition and the structured data table, target template data is determined, and a target poster is generated. On the one hand, by obtaining initial template data of each poster template provided by a target platform, and constructing a structured data table of each initial template data, the efficiency of subsequent target poster template retrieval is facilitated. On the other hand, after receiving poster creation data, based on the template classification requirements of the target object, a plurality of target annotated template data satisfying the template classification requirements of the target object are filtered out from the structured data table. Then, based on the target retrieval keywords corresponding to the main title or text template annotation information provided by the target object, the plurality of target annotated template data are matched to realize accurate filtering of initial target template data. Finally, the target retrieval keywords corresponding to the annotation information are replaced with the target text content corresponding to the annotation information in the initial target template data to obtain target template data, and an intelligent poster is generated.

[0062] In specific embodiments, the initial template data of each poster template stored in the target platform database is acquired, the initial template data is text processed, and the label information corresponding to each target text data is determined to obtain labeled template data, including: The initial template data of each poster template stored in the target platform database is acquired, and the text data in the initial template data is extracted; The text data is sorted according to the position information of the text data in the initial template data; The correlation of each text data is determined according to the semantic information of the sorted text data, and the text data with correlation is text correlated to obtain target text data; The label information of each target text data is determined according to the semantic information of each target text data; The labeled template data is obtained according to the label information of each target text data and each target text data.

[0063] In specific embodiments, the labeled template data is classified according to the label information of the labeled template data, and a structured data table is established, including: The labeled template data is classified according to the template classification label information of the labeled template data to obtain labeled template data of different template classifications; The structured data table of template identification, template architecture label information and target text data is constructed according to the labeled template data included in different template classifications.

[0064] In specific embodiments, the labeled template data is classified according to the template classification label information of the labeled template data to obtain labeled template data of different template classifications, including: According to the template classification label information of the labeled template data, the first level label and the second level label of the template classification corresponding to each labeled template data are determined; According to the first level label and the second level label of the template classification corresponding to each labeled template data, the labeled template data is classified to obtain labeled template data of different template classifications.

[0065] In specific embodiments, in response to receiving the poster creation data submitted by the target object, the retrieval condition is constructed according to the poster creation data, including: In response to receiving the poster creation data submitted by the target object, the poster creation data is processed to obtain a plurality of poster creation word groups; According to the poster creation word group, the target retrieval keyword and the target label information corresponding to the target retrieval keyword are determined; The target retrieval keyword and the target annotation information corresponding to the target retrieval keyword are used to construct a retrieval condition.

[0066] In a specific embodiment, the target retrieval keyword and the target annotation information corresponding to the target retrieval keyword are used to construct a retrieval condition, including: According to the target annotation information corresponding to the target retrieval keyword, the target annotation information corresponding to the target retrieval keyword is classified to obtain target template classification annotation information and target template architecture annotation information; Based on the target template classification annotation information and the target retrieval keyword corresponding to the target template classification annotation information, a first retrieval condition is constructed; Based on the target template architecture annotation information and the target retrieval keyword corresponding to the target template architecture annotation information, a second retrieval condition is constructed based on the first retrieval condition.

[0067] In a specific embodiment, the target template data is determined according to the retrieval condition and the structured data table, and a target poster is generated, including: According to the target template classification annotation information included in the retrieval condition and the target retrieval keyword corresponding to the target template classification annotation information, the target annotation template data is obtained from the structured data table; According to the target template architecture annotation information included in the retrieval condition and the target retrieval keyword corresponding to the target template architecture annotation information, the target text data corresponding to the target template architecture annotation information is filtered from the target annotation template data; According to the matching degree of the target text data corresponding to the target template architecture annotation information and the target retrieval keyword corresponding to the target template architecture annotation information, the initial target template data is determined; According to the initial target template data and the target retrieval keyword corresponding to the target template architecture annotation information, the target template data is determined, and the target poster is generated based on the target template data.

[0068] In a specific embodiment, the target template data is determined according to the initial target template data and the target retrieval keyword corresponding to the target template architecture annotation information, including: Hot information is obtained and analyzed to determine the attribute information corresponding to each hot information; According to the attribute information corresponding to each hot information, the relevance of each hot information to the poster creation data submitted by the target object is determined; According to the relevance of each hot information to the poster creation data submitted by the target object, target hot information with a relevance satisfying a preset relevance is selected; According to the initial target template data, the target search keyword corresponding to the target template architecture annotation information, and the target hot spot information, target template data is determined.

[0069] The embodiment of the present application further provides a computer device, please refer to Figure 3 , Figure 3 The basic structure block diagram of the computer device of the embodiment is shown in the figure.

[0070] The computer device comprises a memory 510 and a processor 520 which are connected to each other in communication through a system bus. It should be noted that only the computer device with components 510-520 is shown in the figure, but it should be understood that all the components shown are not required to be implemented, and more or fewer components can be alternatively implemented. It can be understood by those skilled in the art that the computer device herein is a device capable of automatically performing numerical calculation and / or information processing according to pre-set or stored instructions, and the hardware thereof includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0071] The computer device can be a desktop computer, a notebook computer, a palm computer, a cloud server and the like. The computer device can interact with a user through a keyboard, a mouse, a remote controller, a touchpad, a voice control device and the like.

[0072] The memory 510 includes at least one type of readable storage medium, including non-volatile memory or volatile memory, for example, flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. The RAM can include static RAM or dynamic RAM. In some embodiments, the memory 510 can be an internal storage unit of the computer device, for example, a hard disk or a memory of the computer device. In other embodiments, the memory 510 can also be an external storage device of the computer device, for example, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, or a flash card, etc. equipped on the computer device. Of course, the memory 510 can include both an internal storage unit and an external storage device of the computer device. In this embodiment, the memory 510 is generally used to store an operating system and various application software installed on the computer device, for example, program codes of the above-described method, etc. In addition, the memory 510 can also be used to temporarily store various data that has been output or will be output.

[0073] The processor 520 is generally used to perform the overall operation of the computer device. In this embodiment, the memory 510 is used to store program codes or instructions, including computer operation instructions, and the processor 520 is used to execute the program codes or instructions stored in the memory 510 or process data, for example, run the program codes of the above-described method.

[0074] In this article, the bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus system can be divided into address bus, data bus, control bus, etc. For the convenience of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0075] Another embodiment of the present application also provides a computer readable medium, which can be a computer readable signal medium or a computer readable medium. The processor in the computer reads the computer readable program code stored in the computer readable medium, so that the processor can perform the function actions specified in each step or combination of steps in the above method; generate the device implementing the function actions specified in each block or combination of blocks in the block diagram.

[0076] The computer readable medium includes but is not limited to electronic, magnetic, optical, electromagnetic, infrared, semiconductor system, device or apparatus, or any appropriate combination of the foregoing, for storing program code or instructions, which include computer operation instructions, and processor for executing the program code or instructions of the above method stored in the memory.

[0077] The definition of the memory and the processor can refer to the description of the foregoing computer device embodiment, which will not be repeated here.

[0078] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiment described above is only schematic, for example, the division of modules or units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be through some interface, indirect coupling or communication connection between the devices or units, which can be electrical, mechanical or other forms.

[0079] The function units or modules in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software function unit.

[0080] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the embodiments of the method of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0081] Unless the context clearly indicates otherwise, as used herein and in the appended claims, the singular form "a," "an," and "the" include plural references unless the context clearly dictates otherwise. Accordingly, the use of "a" or "an" herein and in the following claims is intended to be interpreted to include the plural, unless the context clearly indicates otherwise. Similarly, the words "comprise," "comprises," and "comprising" are to be interpreted inclusively rather than exclusively. Likewise, the terms "include," "including," and "includes" should be construed to be inclusive, unless otherwise indicated herein. Where the term "example" is used occurring in this document, particularly with respect to a term or phrase, the "example" is merely an example and is not to be construed as preferred or advantageous over other examples.

[0082] Further aspects and scope of adaptation become apparent from the description provided herein. It should be understood that various aspects of the present application can be practiced alone or in combination with one or more other aspects. It should also be understood that the description and specific examples herein are intended to be illustrative only and are not intended to limit the scope of the present application.

[0083] The above detailed description of several embodiments of the present disclosure, but obviously, those skilled in the art can make various modifications and variations to the embodiments of the present disclosure without departing from the spirit and scope of the present disclosure. The protection scope of the present disclosure is defined by the appended claims.

Claims

1. A method for generating intelligent posters based on a large model, characterized in that, include: Obtain the initial template data of each poster template stored in the target platform's database, perform text processing on the initial template data, and determine the annotation information corresponding to each target text data to obtain annotated template data; Based on the annotation information of the annotation template data, the annotation template data is classified and a structured data table is established, wherein the structured data table includes template identifier, annotation information and target text data; In response to receiving poster creation data submitted by the target object, construct search criteria; Based on the search criteria and structured data table, target template data is determined, and target posters are generated.

2. The method according to claim 1, characterized in that, The process of obtaining initial template data for each poster template stored in the target platform's database, performing text processing on the initial template data, and determining the annotation information corresponding to each target text data to obtain annotated template data includes: Obtain the initial template data of each poster template stored in the target platform's database, and extract the text data from the initial template data; The text data is sorted according to the position information of the text data in the initial template data; Based on the semantic information of each sorted text data, the relationship between each text data is determined, and the text data with the relationship are associated to obtain the target text data; Based on the semantic information of each target text data, determine the annotation information of each target text data; Based on the annotation information of each target text data and each target text data, annotation template data is obtained.

3. The method according to claim 1, characterized in that, The step of classifying the annotation template data according to the annotation information and establishing a structured data table includes: Based on the template classification annotation information of the annotation template data, the annotation template data is classified to obtain annotation template data of different template categories; Based on the labeled template data included in different template categories, construct a structured data table containing template identifiers, template structure annotation information, and target text data.

4. The method according to claim 3, characterized in that, The step of classifying the annotation template data according to the template classification annotation information of the annotation template data to obtain annotation template data of different template categories includes: Based on the template classification annotation information of the labeled template data, determine the primary and secondary labels of the template category corresponding to each labeled template data; Based on the primary and secondary labels of the template category corresponding to each annotation template data, the annotation template data is classified to obtain annotation template data of different template categories.

5. The method according to claim 1, characterized in that, The response to receiving poster creation data submitted by the target object involves constructing search conditions based on the poster creation data, including: In response to receiving poster creation data submitted by the target object, the poster creation data is segmented to obtain multiple poster creation phrases; Based on the poster creation phrases, determine the target search keywords and the target annotation information corresponding to the target search keywords; Construct search criteria based on the target search keywords and the target annotation information corresponding to the target search keywords.

6. The method according to claim 5, characterized in that, The step of constructing search conditions based on target search keywords and the target annotation information corresponding to the target search keywords includes: Based on the target annotation information corresponding to the target search keywords, the target annotation information corresponding to the target search keywords is classified to obtain target template classification annotation information and target template architecture annotation information; Based on the target template classification annotation information and the target search keywords corresponding to the target template classification annotation information, the first search condition is constructed; Based on the target template architecture annotation information and the target search keywords corresponding to the target template architecture annotation information, a second search condition is constructed on the basis of the first search condition.

7. The method according to claim 6, characterized in that, The step of determining the target template data and generating the target poster based on the search criteria and structured data table includes: Based on the target template classification labeling information and the target search keywords corresponding to the target template classification labeling information included in the search conditions, the target labeling template data is obtained from the structured data table; Based on the target template architecture annotation information and the target search keywords corresponding to the target template architecture annotation information included in the search conditions, target text data corresponding to the target template architecture annotation information is filtered from the target annotation template data. The initial target template data is determined based on the matching degree between the target text data corresponding to the target template architecture annotation information and the target search keywords corresponding to the target template architecture annotation information; Based on the initial target template data and the target search keywords corresponding to the target template architecture annotation information, the target template data is determined, and the target poster is generated based on the target template data.

8. The method according to claim 7, characterized in that, The step of determining the target template data based on the initial target template data and the target search keywords corresponding to the target template architecture annotation information includes: Acquire trending information, analyze it, and determine the attribute information corresponding to each trending information. Based on the attribute information corresponding to each hot topic, determine the correlation between each hot topic and the poster creation data submitted by the target object; Based on the correlation between each hot topic and the poster creation data submitted by the target audience, select the target hot topic that meets the preset correlation. The target template data is determined based on the initial target template data, the target search keywords corresponding to the target template architecture annotation information, and the target hotspot information.

9. A smart poster generation device based on a large model, characterized in that, include: The data processing module is used to obtain the initial template data of each poster template stored in the target platform database, perform text processing on the initial template data, and determine the annotation information corresponding to each target text data to obtain annotated template data. The data table construction module is used to classify the annotation template data according to the annotation information of the annotation template data and establish a structured data table, wherein the structured data table includes template identifier, annotation information and target text data; The search criteria construction module is used to construct search criteria in response to receiving poster creation data submitted by the target object; The poster generation module is used to determine the target template data and generate the target poster based on the search conditions and the structured data table.

10. A computer device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 8.

Images