Cloud-based sample pool marketing generation system

Abstract

The application discloses a cloud sample pool-based marketing generation system, comprising a material library receiving module, a product data processing module and a graphic planning module, characterized in that: the output end of the material library receiving module is connected with the input end of the product data processing module through a small program interface, the output end of the product data processing module is electrically connected with the input end of the graphic planning module through a cloud storage interface, the material library receiving module is used for receiving various format files uploaded by users, the product data processing module is used for processing product data by parameters transmitted through the front end and the back end of an application, and the graphic planning module is used for planning display arrangement of marketing results, and the application has the characteristics of high generation quality and good marketing effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent marketing, in particular to a cloud sample pool-based marketing generation system. BACKGROUND

[0002] With the development of society, browsing websites and social media through computers, mobile phones and other mobile terminals has become a way of life for everyone. A large amount of information data is flooding people's lives, bringing great opportunities to internet marketing. Traditional marketing methods are paper marketing, such as paper publishing media like newspapers and posters. Paper media marketing has unparalleled advantages in design and visual presentation, such as no deviation in typesetting and color. However, with the development of the times, the disadvantages of paper media marketing have gradually increased, such as high printing cost, low timeliness, limited marketing effect, low marketing accuracy, and in the field of visual design, designers often spend a considerable amount of time and labor for some simple requirements, resulting in resource waste and low efficiency.

[0003] In the future, precision marketing is a trend. Under the background of big data and large flow, the display effect of poster resources requires high quality. Therefore, it is necessary to design a cloud sample pool-based marketing generation system with high quality and good marketing effect. SUMMARY

[0004] The present application relates to the technical field of intelligent marketing, in particular to a cloud sample pool-based marketing generation system.

[0005] To solve the above technical problems, the present application provides the following technical solution: a cloud sample pool-based marketing generation system, comprising a material library receiving module, a product data processing module and a graphic planning module. The output end of the material library receiving module is connected to the input end of the product data processing module through a small program interface. The output end of the product data processing module is connected to the input end of the graphic planning module through a cloud storage interface. The material library receiving module is used to receive a variety of format files uploaded by users. The product data processing module is used to process product data through parameters transmitted by the front and back ends of the application. The graphic planning module is used to plan the display arrangement of marketing results.

[0006] According to the above technical solution, the material library receiving module comprises an API interface module, a source file uploading module and a feature extraction module. The interface end of the API interface module is connected to the input end of the source file uploading module. The interface end of the API interface module is connected to the input end of the feature extraction module. The API interface module is used to provide an interface for file transmission and information interaction. The source file uploading module is used to provide a demand file uploading channel for users. The feature extraction module is used to extract feature sequences in the files uploaded by users.

[0007] According to the technical scheme, the product data processing module comprises a product data acquisition module, a data screening module, a parameter judgment module and a marking module, the output end of the product data acquisition module is electrically connected with the input end of the data screening module, the output end of the data screening module is electrically connected with the input end of the parameter judgment module, and the output end of the parameter judgment module is electrically connected with the input end of the marking module, the product data acquisition module is used for acquiring product data through an interface of a mini program, the data screening module is used for screening products according to data parameters transmitted by a front end, the parameter judgment module is used for judging parameters of product data cyclically, and the marking module is used for marking video pictures that need to be further operated.

[0008] According to the technical scheme, the graphic planning module comprises a GAN image generation training module, a sequence generation model fitting module, a comprehensive optimization module and a back-end connection module, the output end of the GAN image generation training module is electrically connected with the sequence generation model fitting module, the output end of the sequence generation model fitting module is electrically connected with the input end of the back-end connection module, the GAN image generation training module is used for end-to-end image model generation training, the sequence generation model fitting module is used for fitting a process of stacking materials according to a sequence generation model, the comprehensive optimization module is used for performing optimization operation on graphics by using a target function, and the back-end connection module is used for performing localization operation on the generated optimized graphics.

[0009] According to the technical scheme, the operation method of the cloud sample pool-based marketing generation system comprises the following steps:

[0010] Step S1: The system is connected to a mini program and an APP, a user registers personal or company information in the mini program and the APP, and uploads source files in multiple formats through a source file uploading module;

[0011] Step S2: The format of the source file uploaded by the user is judged, a corresponding operation mode is selected, and feature information in the source file is extracted as product data;

[0012] Step S3: The extracted product data is processed, and the processing result is fed back to the registered user;

[0013] Step S4: Image planning is performed to complete generation of a marketing product.

[0014] According to the technical scheme, in steps S1-S2, the operation process of the user and the feedback method of the system further comprise the following steps:

[0015] Step A: the user registers and logs in through the applet or APP, and uploads the demand source file; the user uploads the demand source file and submits the demand note information, so that the system can more intuitively and conveniently obtain the user's preferences and improve the marketing effect;

[0016] Step B: analyze and judge the source file uploaded by the user, and select the opening mode according to different formats to read the file content; the user can submit PDF, picture compression package, video and other files, and different opening modes need to be used for different files to improve the compatibility of the system;

[0017] Step C: combine the user's demand with the content of the uploaded source file to extract the feature sequence contained in the file. Different image styles can extract corresponding feature sequences, which can more significantly display the user demand style, such as color, texture, main body, arrangement and other features. These abstract features are extracted in the form of feature sequence to increase the stereoscopic degree of user demand, and the system can more accurately grasp the user demand;

[0018] According to the above technical scheme, the step S3 further includes the following steps:

[0019] Step S31: open the applet or H5 platform to obtain the uploaded product data; using the applet or H5 platform can obtain the data more directly and facilitate debugging;

[0020] Step S32: select parameter set according to the storage order to filter data and set return value; after data filtering, the output parameter needs to be valid and there needs to be a storage address, and setting the return value can avoid the parameter illegal caused by invalid data in the output process;

[0021] Step S33: loop the operation of step S32 to perform parameter judgment and provide user viewing permission and group binding suggestion; loop the product data to perform parameter judgment to determine whether the user needs to view and whether the known group needs to be bound;

[0022] Step S34: according to the judgment result of the front end, judge whether the product is bound to the sample, and display the corresponding icon, and mark the product sample that meets the recommendation requirements. The user product that has completed the judgment will present the marketing value, the front end program automatically judges whether it meets the recommendation requirements, and forms a good marketing effect;

[0023] According to the above technical scheme, in step S4, the image planning method includes the following steps:

[0024] Step S41: create a canvas instance, add a rectangular background and generate a GAN image training program; the GAN image training program can realize an end-to-end image generation model, and is combined into the canvas instance to perform normalization processing on the original image and the feature value;

[0025] Step S42: sequence generation model fitting is performed, and the feature values extracted from the source file are sequentially stacked; after the main body information is input, the extracted user demand feature value sequence is fitted with the main body information, so that the material coating is stacked;

[0026] Step S43: the image is segmented, morphological conversion is performed, a three-value Trimap is generated, and a feature vector of user demand is established Wherein x i , y i , z i represent different feature values of the image; the user's demand is represented in the form of a feature vector, which facilitates digital comparison of the implementation results with the user's demand, and the results are more intuitive;

[0027] Step S44: the actual feature vector of the generated result is extracted and compared with the user demand to determine the matching result.

[0028] According to the above technical scheme, in step S44, after the user demand is compared, the matching result is determined, and the method for calculating the matching degree H is:

[0029] The feature vector of the user demand and the actual feature vector of the generated result are similar, the matching result is fed back, when the feature vector of the user demand and the actual feature vector are equal in size and direction, the matching degree is 1, when the feature vector of the user demand and the actual feature vector are perpendicular in direction, the matching degree is 0, therefore the cosine value between the vectors is used to represent the matching degree.

[0030] According to the above technical scheme, in step S44, the calculation formula of the matching degree H is:

[0031]

[0032] In the formula, the value of H is related to the matching degree of the feature vector of the user demand and the actual feature vector of the generated result , the closer H is to 1, the higher the matching degree of the user demand generated result is, and the user's satisfaction degree is also higher.

[0033] Compared with the prior art, the present application has the beneficial effects that: the present application,

[0034] (1) by setting with the material library receiving module, while providing the user end source file upload module, through the interface provided by the API interface module, the user provides source file upload and demand upload, meet the user demand;

[0035] (2) by setting with the feature extraction module, using big data technology to extract and integrate the features of the uploaded source files and demands, and fitting the demand feature vector of the user;

[0036] (3) by setting with product data processing module, further processing product data, and through the small program or APP for data screening and parameter judgment, and marking;

[0037] (4) by setting with the graphic planning module, the obtained material is based on the feature vector graphic planning;

[0038] (5) by setting with the GAN image generation training module, realize the end-to-end image generation model, and combine to the canvas instance, carry out the normalization processing of original image and characteristic value;

[0039] (6) by setting with the sequence generation model fitting module, according to the extracted user demand characteristic value sequence, and the main information fitting, realize the layering of material coating, and express the user's demand in the form of feature vector, convenient to realize the digital comparison between the achievements and the user demand, the result is more intuitive. BRIEF DESCRIPTION OF DRAWINGS

[0040] The accompanying drawings are used to provide further understanding of the present application, and constitute a part of the specification, and are used to explain the present application together with embodiments of the present application, and do not constitute a limitation to the present application. In the drawings:

[0041] Figure 1 is the system module composition schematic diagram of the present application. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0043] Please refer to Figure 1The application provides a technical scheme: a cloud sample pool-based marketing generation system, comprising a material library receiving module, a product data processing module and a graphic planning module, the output end of the material library receiving module is connected with the input end of the product data processing module through a mini-program interface, the output end of the product data processing module is connected with the input end of the graphic planning module through a cloud storage interface, the material library receiving module is used for receiving a plurality of format files uploaded by a user, the product data processing module is used for processing product data through parameters transmitted by a front-end and a back-end, and the graphic planning module is used for planning display and arrangement of marketing results.

[0044] The material library receiving module comprises an API interface module, a source file uploading module and a feature extraction module, the interface end of the API interface module is connected with the input end of the source file uploading module, the interface end of the API interface module is connected with the input end of the feature extraction module, the API interface module is used for providing an interface for file transmission and information interaction, the source file uploading module is used for providing a demand file uploading channel for the user, and the feature extraction module is used for extracting a feature sequence in the file uploaded by the user.

[0045] The product data processing module comprises a product data acquisition module, a data screening module, a parameter judgment module and a marking module, the output end of the product data acquisition module is connected with the input end of the data screening module, the output end of the data screening module is connected with the input end of the parameter judgment module, the output end of the parameter judgment module is connected with the input end of the marking module, the product data acquisition module is used for acquiring product data through a mini-program interface, the data screening module is used for screening products according to data parameters transmitted by a front-end, the parameter judgment module is used for judging parameters of product data in a loop, and the marking module is used for marking a video picture that needs to be further operated.

[0046] The graphic planning module comprises a GAN image generation training module, a sequence generation model fitting module, a comprehensive optimization module and a back-end connection module, the output end of the GAN image generation training module is connected with the sequence generation model fitting module, the output end of the sequence generation model fitting module is connected with the input end of the back-end connection module, the GAN image generation training module is used for end-to-end image model generation training, the sequence generation model fitting module is used for fitting a material stacking process according to a sequence generation model, the comprehensive optimization module is used for optimizing a graphic by using a target function, and the back-end connection module is used for localizing the generated optimized graphic.

[0047] The operation method of the cloud sample pool-based marketing generation system comprises the following steps:

[0048] Step S1: The system is connected to a mini-program and an APP, a user registers personal or company information in the mini-program and the APP, and uploads a plurality of format source files through a source file uploading module;

[0049] Step S2: judging the format of the source file uploaded by the user, selecting the corresponding operation mode, and extracting the feature information in the source file as product data;

[0050] Step S3: processing the extracted product data and feeding back the processing result to the registered user;

[0051] Step S4: image planning is performed to complete the generation of marketing products.

[0052] In steps S1-S2, the operation flow of the user and the feedback method of the system further include the following steps:

[0053] Step A: the user registers and logs in through the applet or APP, and uploads the demand source file; the user uploads the demand source file and submits the demand note information, so that the system can more directly and conveniently obtain the user's preferences and improve the marketing effect;

[0054] Step B: analyzing and judging the source file uploaded by the user, and selecting the opening mode according to different formats to read the file content; the user can submit PDF, picture compression package, video and other files, and different opening modes need to be used for different files to improve the compatibility of the system;

[0055] Step C: combining the user's demand and the content of the uploaded source file, the feature sequence contained in the file is extracted. Different image styles can extract corresponding feature sequences, which can more significantly display the user demand style, such as color, texture, main body, arrangement and other features. These abstract features are extracted in the form of feature sequence to increase the stereoscopic degree of user demand, and the system can more accurately grasp the user demand;

[0056] Step S3 further includes the following steps:

[0057] Step S31: opening the applet or H5 platform to obtain the uploaded product data; using the applet or H5 platform can more directly obtain the data and is convenient for debugging;

[0058] Step S32: selecting parameter set according to the storage order to filter data and setting return value; after data filtering is completed, the output parameter needs to be valid and there needs to be a storage address, and setting the return value can avoid the parameter illegal caused by invalid data in the output process;

[0059] Step S33: repeating the operation of step S32 to judge the parameters and provide user viewing permission and group binding suggestions; the product data is cycled to judge the parameters, determine whether the registered user needs to view, and judge whether to bind the known group;

[0060] Step S34: According to the front-end judgment result, judge whether the product is bound to the sample, and display the corresponding icon, and mark the product sample that meets the recommendation requirements. The user product that has completed the judgment will present the marketing value, and the front-end program automatically judges whether it meets the recommendation requirements, forming a good marketing effect;

[0061] In step S4, the method of image planning includes the following steps:

[0062] Step S41: Create a canvas instance, add a rectangular background and generate a GAN image training program; the GAN image training program can realize an end-to-end image generation model, and is combined into the canvas instance to perform normalization processing on the original image and the feature value;

[0063] Step S42: Perform sequence generation model fitting, and sequentially stack the feature values extracted from the source file; after inputting the main body information, the extracted user demand feature value sequence is fitted with the main body information to realize the layering of the material coating;

[0064] Step S43: Segment the image, perform morphological conversion, generate a three-value Trimap, and establish a feature vector of user demand Where x i , y i , and z i represent different feature values of the image; the user's demand is represented in the form of a feature vector, which facilitates digital comparison of the implementation results with user demand, and the results are more intuitive;

[0065] Step S44: Extract the actual feature vector of the generated result and compare it with the user demand to judge the matching result.

[0066] In step S44, after comparing with the user demand, the method for judging the matching result and calculating the matching degree H is:

[0067] The feature vector of the user demand and the feature vector of the generated result are compared to feedback the matching result. When the feature vector of the user demand is equal in size and direction to the actual feature vector , the matching degree is 1, and when the feature vector of the user demand is perpendicular to the actual feature vector , the matching degree is 0, so the cosine value between the vectors is used to represent the matching degree.

[0068] In step S44, the calculation formula of the matching degree H is:

[0069]

[0070] wherein the value of H is related to the matching degree between the actual feature vector of the generated result and the feature vector of the user demand The closer H is to 1, the higher the matching degree of the generated result to the user demand, and the higher the satisfaction of the user.

[0071] It should be noted that the relational terms herein such as first and second and the like are used solely to distinguish one from another entity or action, without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0072] Finally, it should be noted that the above-mentioned only constitutes the preferred embodiments of the present application, and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art will still be able to modify the technical solutions described in the foregoing embodiments, or make equivalent replacements to some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.​

Claims

1. A cloud-based sample pool marketing generation system, comprising a material library receiving module, a product data processing module, and a graphic planning module, characterized in that: The output end of the material library receiving module is connected with the input end of the product data processing module through a small program interface, the output end of the product data processing module is connected with the input end of the graphic planning module through a cloud storage interface, the material library receiving module is used for receiving a user uploaded multiple format files, the product data processing module is used for processing product data through parameters transmitted by the front and rear ends, and the graphic planning module is used for planning display arrangement of marketing results. The operation method of the cloud sample pool marketing generation system comprises the following steps: Step S1: connecting the system to a small program and an APP, registering personal or company information in the small program and the APP, and uploading multiple format source files through a source file uploading module; Step S2: judging the format of the source file uploaded by the user, selecting a corresponding operation mode, and extracting feature information in the source file as product data; Step S3: processing the extracted product data and feeding back the processing result to the registered user; Step S4: image planning and completing generation of a marketing product; In step S4, the image planning method comprises the following steps: Step S41: creating a canvas instance, adding a rectangular background and generating a GAN image training program; Step S42: performing sequence generation model fitting and sequentially stacking the extracted feature values in the source file; Step S43: segmenting the image, performing morphological conversion, generating a three-value Trimap, and establishing a feature vector of user demand wherein , represent different feature values of the image; Step S44: extracting the actual feature vector of the generated result and comparing with the user demand to judge the matching result; In the step S44, after comparing with the user demand, it is judged whether the matching result matches the calculation matching degree The method is as follows: The feature vector of the user demand The actual feature vector of the generated result The feedback matching result of the group similarity, when the feature vector of the user demand The actual feature vector The matching degree is 1 when the size and direction are the same, and the matching degree is 0 when the feature vector of the user demand The actual feature vector The direction is perpendicular, so the cosine value between the vectors is used to represent the matching degree, and the calculation formula of the matching degree is: ； In the formula, The value and the feature vector of user needs The actual feature vector of the generated result The degree of matching is related to the H value. The closer H is to 1, the higher the degree of matching between the user's needs and the generated results, and the higher the user's satisfaction will be.

2. The cloud-based sample pool marketing generation system of claim 1, wherein: The material library receiving module comprises an API interface module, a source file uploading module and a feature extraction module, the interface end of the API interface module is connected with the input end of the source file uploading module, the interface end of the API interface module is connected with the input end of the feature extraction module, the API interface module is used for providing a file transmission and information interaction interface, the source file uploading module is used for providing a demand file uploading channel for a user, and the feature extraction module is used for extracting a feature sequence in a file uploaded by a user.

3. The cloud-based sample pool marketing generation system of claim 2, wherein: The product data processing module comprises a product data acquisition module, a data screening module, a parameter judgment module and a marking module, the output end of the product data acquisition module is connected with the input end of the data screening module, the output end of the data screening module is connected with the input end of the parameter judgment module, the output end of the parameter judgment module is connected with the input end of the marking module, the product data acquisition module is used for acquiring product data through a small program, the data screening module is used for screening products according to data parameters transmitted by the front end, the parameter judgment module is used for judging parameters of product data in cycles, and the marking module is used for marking video pictures that need to be further operated.

4. The cloud-based sample pool marketing generation system of claim 3, wherein: The graphic planning module includes a GAN image generation training module, a sequence generation model fitting module, a comprehensive optimization module and a backend connection module, the output end of the GAN image generation training module is electrically connected with the sequence generation model fitting module, the output end of the sequence generation model fitting module is electrically connected with the input end of the backend connection module, the GAN image generation training module is used for end-to-end image model generation training, the sequence generation model fitting module is used for fitting the process of material stacking according to the sequence generation model, the comprehensive optimization module is used for optimizing the graphics by using the target function, and the backend connection module is used for localizing the generated optimized graphics.

5. The cloud-based sample pool marketing generation system of claims 1-4, wherein: In the steps S1-S2, the operation process of the user and the feedback method of the system further include the following steps: Step A: the user registers and logs in through the applet or APP, and uploads the demand source file; Step B: analyze and judge the source file uploaded by the user, and select the opening mode according to different formats to read the file content; Step C: combining the user's demand with the content of the uploaded source file, extracting the feature sequence contained in the file.

6. The cloud-based sample pool marketing generation system of claim 5, wherein: The step S3 further includes the following steps: Step S31: open the applet or H5 platform, and obtain the uploaded product data; Step S32: select the parameter set according to the storage order to filter the data, and set the return value; Step S33: cycle the operation of step S32, judge the parameters, and provide user viewing permission and group binding suggestions; Step S34: according to the front-end judgment result, judge whether the product is bound to the sample, and display the corresponding icon, and mark the product sample that meets the recommendation requirements.

Images