A natural language-based enterprise user template production tool and method

Abstract

The present application provides a natural language-based enterprise user template production tool and method, which relate to the field of information processing technology. The tool and method receive enterprise users' template generation demand information; analyze store template design requirements based on an NLP model to determine demand feature information; perform template registration on the demand feature information based on template registration recognition multi-channel; perform rendering optimization based on a store product image set and a predetermined product image background set; then, perform animation recommendation comparison on the enterprise users to generate user store templates; and perform correlation adjustment on the user store templates based on real-time store monitoring data, thereby solving the technical problems of poor market response immediacy of user templates, low design accuracy and slow response speed, and achieving the goal of improving the degree of automation and personalization accuracy of template generation, so that the template can reflect the latest status of store operations and market changes in real time, thereby improving efficiency while reducing market investment costs.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a natural language-based enterprise user template production tool and method. Background Art

[0002] With the rapid development of digital marketing and e-commerce, personalized and automated store template design has become an important tool for corporate users to pursue brand uniqueness and enhance user experience. However, with the intensification of market competition and the diversification of consumer demand, traditional template design methods face many challenges. These methods often rely on manual operations, which are not only time-consuming and labor-intensive, but also difficult to update in real time to match the latest market trends and users' personalized needs. Moreover, in the existing technology, the template design process is mostly static, lacking flexibility and adaptability, especially when dealing with a large number of complex design requirements, such as maintaining brand consistency on multiple pages or quickly adjusting the design according to seasons and promotions. Existing methods are often slow to respond, low in accuracy and low in efficiency. These limitations make it difficult for companies to fully utilize digital tools to optimize the performance of their online stores, which may damage brand image and sales performance.

[0003] In summary, the existing technology has the following technical problems: poor immediacy of user template market response, low design accuracy and slow response speed, and inability to dynamically adjust template personalization in real time, resulting in low user output efficiency and high cost. Summary of the invention

[0004] The present application provides a natural language-based enterprise user template production tool and method, which is used to solve the technical problems in the prior art of poor market response immediacy of user templates, low design accuracy and slow response speed, and inability to perform real-time dynamic template personalization adjustment, resulting in low user output efficiency and high cost.

[0005] In view of the above problems, the present application provides a natural language-based enterprise user template production tool and method.

[0006] In a first aspect of the present application, a natural language-based enterprise user template production tool is provided, the tool comprising:

[0007] An enterprise user interaction module, the enterprise user interaction module is used to receive template generation demand information of an enterprise user, wherein the template generation demand information includes the store template design requirements, store product image set and predetermined product image background set of the enterprise user; a natural language parsing module, the natural language parsing module is used to parse the store template design requirements based on the NLP model and determine the store template demand feature information; a store template registration module, the store template registration module is used to perform template registration on the store template demand feature information based on template registration recognition multi-channel to construct a first store template set; a store template optimization module, the store template optimization module is used to perform rendering optimization on the first store template set based on the store product image set and the predetermined product image background set to generate a second store template set; an animation recommendation comparison module, the animation recommendation comparison module is used to perform animation recommendation comparison on the enterprise user based on the second store template set to generate a user store template; a template dynamic adjustment module, the template dynamic adjustment module is used to obtain the real-time store monitoring data of the enterprise user, and make associated adjustments to the user store template according to the real-time store monitoring data.

[0008] A second aspect of the present application provides a method for producing an enterprise user template based on natural language, the method comprising:

[0009] Receive template generation requirement information from an enterprise user, wherein the template generation requirement information includes the enterprise user's store template design requirement, store product image set and predetermined product image background set; parse the store template design requirement based on an NLP model to determine store template requirement feature information; perform template registration on the store template requirement feature information based on template registration recognition multi-channel to construct a first store template set; perform rendering optimization on the first store template set based on the store product image set and the predetermined product image background set to generate a second store template set; perform animation recommendation comparison on the enterprise user based on the second store template set to generate a user store template; obtain real-time store monitoring data of the enterprise user, and make correlation adjustments to the user store template according to the real-time store monitoring data.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The present application provides a natural language-based enterprise user template production tool, which receives template generation demand information from enterprise users, wherein the template generation demand information includes the enterprise user's store template design requirements, store product image sets and predetermined product image background sets; analyzes the store template design requirements based on an NLP model to determine store template demand feature information; performs template registration on the store template demand feature information based on template registration recognition multi-channel to construct a first store template set; performs rendering optimization on the first store template set based on the store product image set and the predetermined product image background set to generate a second store template set; and generates a store template set based on the second store template set. The shop template set performs animation recommendation comparison on the enterprise user to generate a user shop template; obtains the real-time shop monitoring data of the enterprise user, and makes associated adjustments to the user shop template according to the real-time shop monitoring data, thereby solving the technical problems of poor market response immediacy of user templates, low design accuracy and slow response speed, and inability to perform real-time dynamic personalized adjustment of templates, resulting in low user output efficiency and high costs. It achieves the goal of improving the degree of automation and personalized accuracy of template generation, while increasing the dynamic adjustment capability of templates, so that templates can reflect the latest status of store operations and market changes in real time, achieving the technical effect of improving efficiency while reducing market investment costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic diagram of the structure of a natural language-based enterprise user template production tool provided in an embodiment of the present application.

[0013] Figure 2 A flowchart of a natural language-based enterprise user template production method provided in an embodiment of the present application.

[0014] Explanation of the accompanying drawings: enterprise user interaction module 11, natural language analysis module 12, store template matching module 13, store template optimization module 14, animation recommendation comparison module 15, template dynamic adjustment module 16. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0016] Embodiment 1, as Figure 1 As shown, the present application provides a natural language-based enterprise user template production tool, the tool comprising:

[0017] The enterprise user interaction module 11 is used to receive template generation requirement information of the enterprise user, wherein the template generation requirement information includes the store template design requirement, store product image set and predetermined product image background set of the enterprise user.

[0018] The natural language-based enterprise user template production tool is a system tool that uses natural language processing (NLP) technology to automatically parse and realize the template design needs of enterprise users. This tool automatically generates or adjusts online stores, advertisements or other types of commercial templates to meet the user's specific needs and brand style by understanding the needs expressed by the user in text form. The core advantage of this tool is to improve the automation and response speed of the template design process while ensuring the personalization and consistency of the design.

[0019] The natural language-based enterprise user template production tool includes six modules: enterprise user interaction module, natural language parsing module, store template registration module, store template optimization module, animation recommendation comparison module, and template dynamic adjustment module. Among them, the main function of the enterprise user interaction module is to receive and process template generation demand information from enterprise users. This module is designed as a front-end interface, through which users can directly input or upload their store template design requirements and related image materials.

[0020] Specifically, the enterprise user interaction module mainly includes three functional subcomponents, namely, the demand receiving subcomponent, the image collection subcomponent, and the background setting subcomponent. The demand receiving subcomponent is responsible for collecting the text information input by the user, which describes the user's specific requirements for the store template, such as the style, color scheme, layout preference, etc. of the template. The image collection subcomponent allows users to upload image sets of store products through this subcomponent. These image sets are an indispensable part of template design because they are directly related to the visual performance of the final template. The background setting subcomponent can specify or upload a predetermined product image background set. The selection and application of the background set is an important factor affecting user perception in the design process, which can enhance the visual effect of product display. In the specific operation process, the enterprise user first inputs the design requirements of its store template through the demand receiving subcomponent. For example, a coffee shop may need a template with a modern and warm atmosphere. The user will describe this requirement in simple natural language, such as "I need a modern coffee shop template with warm colors as the main color and a simple and clear layout." Next, the user uploads high-definition pictures of its main products through the image collection subcomponent. These pictures will be used to display in the template to attract customers' attention. Finally, the user selects a suitable background image set in the Background Settings subcomponent, or uploads a specific background design. For example, the user might choose a background with a wooden texture to match the cozy theme of the coffee shop.

[0021] Through the enterprise user interaction module, user needs and related image data can be effectively collected and organized, providing accurate input data for the subsequent natural language parsing module and template design optimization module. This process not only improves the efficiency and accuracy of template design, but also ensures that the final output can accurately meet user expectations and market needs.

[0022] The natural language parsing module 12 is used to parse the store template design requirements based on the NLP model to determine the store template requirement feature information.

[0023] The natural language parsing module is a key component for parsing the store template design requirements from enterprise users. The main function of this module is to use natural language processing (NLP) technology to extract key design requirement feature information from the natural language text submitted by users. This information will then be used to guide the creation and optimization of templates. The natural language parsing module mainly includes the functions of text reception, NLP model processing and feature information determination. Specifically, this module first receives the natural language text transmitted from the enterprise user interaction module. These texts contain the user's specific design requirements for the store template. After that, the NLP model is applied to deeply parse the received text to identify and extract key words and phrases. These key words and phrases express the user's specific requirements for the template, such as style, color, layout, etc. Next, the parsed key words and phrases are converted into feature information of design requirements. These feature information exists in a structured form to facilitate the use of subsequent modules.

[0024] For example, the owner of a jewelry store submits the following requirement through the enterprise user interaction module: "I want an elegant and modern store template that mainly displays diamond and platinum jewelry, and the background color should be elegant gray or white." Then, after the natural language parsing module receives the above text, the NLP model starts to work and recognizes key words such as "elegant", "modern", "diamond", "platinum jewelry", "background color", "gray", "white", etc. Subsequently, these keywords are converted into specific feature information. For example, the style is determined to be elegant and modern; the main products are determined to be diamond and platinum jewelry; and the color scheme is determined to be gray and white. These feature information will be encoded into a data format that can be further processed by the template design system.

[0025] The natural language parsing module can automatically and accurately extract key design requirements from the natural language descriptions of enterprise users, greatly improving the efficiency and accuracy of template design. In addition, the module supports highly personalized template generation, so that the final product can better reflect the user's brand characteristics and market positioning, thereby improving user satisfaction and market competitiveness.

[0026] The store template registration module 13 is used to perform template registration on the store template demand feature information based on template registration recognition multiple channels to construct a first store template set.

[0027] The store template registration module is a core component designed to accurately register the store template requirement feature information based on multi-channel recognition technology and build a preliminary store template set that meets the needs of enterprise users. This module aims to optimize the template selection process and ensure that the selected template meets the specific needs of users to the greatest extent. This module uses multiple data recognition channels to analyze and match the store template requirement features, including style matching, color matching, layout matching, etc. Then, based on the recognition and matching results, a set of preliminary store templates are screened and combined from the huge template database, which have the highest matching degree with user needs. Taking the jewelry store template requirements as an example, user requirements may include "elegant and modern style, mainly displaying diamond and platinum jewelry, with gray or white background color". After receiving these feature information converted by the natural language parsing module, the store template registration module starts recognition and matching through the recognition channel. Specifically, the style matching channel can be used to analyze which templates in the existing template library meet the "elegant and modern" style, which may involve the evaluation of the visual elements and overall design language of the template; the product display matching channel can be used to ensure that the selected template has an appropriate display area and design layout to highlight "diamond and platinum jewelry"; the color matching channel can be used to filter out templates with a main color of "gray or white" to match the user's specific requirements for the background color. After the above multi-channel matching, the module will filter out several templates that best meet the above standards from the entire library. These templates constitute the first store template set, ready for subsequent optimization and final confirmation by the user.

[0028] The store template matching module significantly improves the automation and accuracy of template selection, ensuring that the final template design is closer to the user's actual needs. The multi-channel matching of this module makes template selection not only fast but also efficient, greatly saving design time and cost, while increasing user satisfaction.

[0029] The store template optimization module 14 is used to perform rendering optimization on the first store template set based on the store product image set and the predetermined product image background set to generate a second store template set.

[0030] The store template optimization module is responsible for further refining and optimizing the selected store template. This module mainly performs advanced rendering and visual optimization on the first store template set to ensure that the final template is more attractive to target customers and better showcases the company's products. Advanced image processing technology is used to beautify and unify the style of product images in the template to enhance visual appeal; at the same time, the template background is adjusted according to the predetermined product image background set to match the product style and enhance overall visual coordination. After rendering and background optimization, a new set of optimized store templates is output as the second store template set for users to finally select and use.

[0031] Specifically, assume that the first store template set includes a template designed for a jewelry store, which contains multiple product images of diamond necklaces. During the image rendering optimization process, the module will apply high dynamic range (HDR) processing and color correction technology to ensure that the luster and details of the diamonds are accurately presented, while adjusting the brightness and contrast of the image to make it more vivid and eye-catching. For the same jewelry store template, the background integration optimization function will take into account the requirements of the store owner, such as using a soft gray gradient background to set off the beauty of the diamond. The module uses an intelligent matching algorithm to select the background pattern that best coordinates with the product style and color, and can even perform background blur processing to ensure that the product can become the visual focus. Finally, after completing the optimization of the image and background, the store template optimization module integrates these improved templates into the second store template set, which is not only more visually perfect, but also more functionally in line with the user's marketing and display needs.

[0032] The store template optimization module can significantly improve the visual quality and market appeal of the template. The optimized template can better display the company's products and improve the customer's shopping experience, which directly affects the store's conversion rate and customer satisfaction. In addition, the intelligent optimization processing of this module reduces the time and resource investment of enterprises in manually adjusting templates, and improves the efficiency of the entire template design process.

[0033] The animation recommendation comparison module 15 is used to perform animation recommendation comparison on the enterprise user based on the second store template set to generate a user store template.

[0034] The animation recommendation and comparison module is specially designed to recommend and compare the most suitable animation effects for enterprise users from the optimized second store template set, so as to generate the final user store template that meets the needs and tastes of users. Specifically, suitable animation effects are recommended based on the template characteristics in the second store template set to enhance the attractiveness and interactivity of the template. Then, the matching degree of different animation effects with user needs is compared to ensure that the animation that best highlights the product features and user brand is selected. Finally, the final user store template, including the selected animation effects, is generated based on the user's preferences and comparison results.

[0035] Here we still take the jewelry store as an example. Considering that the second store template set contains templates optimized for jewelry stores, the animation recommendation comparison module will recommend animations such as diamond brilliance and exquisite jewelry rotation display from a variety of animation effects. These animations are designed to enhance the visual appeal of jewelry and make online stores more eye-catching. The module automatically compares the visual effects of various recommended animations with the specific needs of users, such as the speed, style and interaction method of the animation. For example, if the user prefers a low-key and elegant display, the module may recommend a simpler and slower animation effect instead of a fast flashing animation. Based on the comparison results and user feedback, the animation recommendation comparison module will select the best animation effect to integrate into the final store template. For example, if the user selects the diamond brilliance animation and confirms that it is in line with their brand image and marketing goals, the final template will include this animation effect, thereby generating a customized user store template.

[0036] The animation recommendation comparison module improves the visual appeal and user interactivity of the template, helping companies to better showcase their products and enhance consumers' shopping experience. In addition, through intelligent comparison and personalized recommendations, the module ensures the consistency of animation effects with corporate brands and market positioning, thereby improving marketing effectiveness and customer satisfaction.

[0037] The template dynamic adjustment module 16 is used to obtain the real-time store monitoring data of the enterprise user and make correlation adjustments to the user store template according to the real-time store monitoring data.

[0038] The template dynamic adjustment module is used to make intelligent and dynamic adjustments to the user's store templates based on real-time store monitoring data. This module can ensure that the store templates always reflect the latest operational data and market trends, and improve user experience and sales efficiency. Specifically, this module automatically collects real-time store monitoring data from the user's e-commerce platform, such as visitor traffic, user interaction (such as click and scroll behavior), sales conversion rate, etc. Further, the collected data is analyzed and algorithms are used to determine which template features (such as layout, color scheme, animation effects of specific elements, etc.) need to be adjusted to optimize user experience and sales performance. Finally, the template is automatically adjusted based on the analysis results. The adjustments may include changing the layout, updating the color configuration, modifying animation or interactive elements, etc.

[0039] Suppose an online clothing retailer experiences a surge in visitors during a specific promotional event. The template dynamic adjustment module collects and monitors this data in real time, such as the number of page visits and the time users spend on each product page. After analyzing the collected data, the module identifies that a specific product page has low user interaction despite high visits. The analysis may indicate that the page layout or product display needs to be optimized to increase user interaction and purchase conversion rate. Finally, based on the analysis results, the module automatically adjusts the affected product page templates. For example, adding a more prominent call to action button, improving the display of product images, or introducing more attractive product descriptions and optimizing page loading speed. These adjustments take effect immediately without manual intervention.

[0040] The template dynamic adjustment module significantly improves the adaptability of store templates and their ability to respond to market changes, which not only optimizes the user's shopping experience, but also adjusts strategies based on real-time data to maximize sales potential. In addition, the module's automatic adjustment function reduces maintenance costs and improves operational efficiency.

[0041] Furthermore, based on the template registration recognition multi-channel, the template registration is performed on the store template demand feature information to construct a first store template set, and the store template registration module is further used for:

[0042] Activate the online store template library in the store template registration module.

[0043] The online store template library is traversed to extract a first online store template.

[0044] A first shop template feature vector is constructed according to the first online shop template.

[0045] A first template demand feature vector is constructed according to the store template demand feature information.

[0046] Based on the first store template feature vector and the first template demand feature vector, a first template registration recognition coefficient is obtained according to the template registration recognition multi-channel.

[0047] It is determined whether the first template registration recognition coefficient is greater than / equal to a predetermined template registration recognition coefficient.

[0048] If the first template registration recognition coefficient is greater than / equal to the predetermined template registration recognition coefficient, the first online store template is added to the first store template set.

[0049] Specifically, the store template registration module has an online store template library, which is activated when triggered. This library contains a variety of pre-designed store templates, each with different visual styles, layouts, and functional elements to meet the needs of different industries and brands. Then, the module starts to traverse these templates and extract each template information one by one for subsequent feature analysis and registration. For each extracted store template, i.e., the first online store template, a corresponding feature vector is constructed as the first store template feature vector. This feature vector contains all the key visual and functional attributes of the template, such as color scheme, layout structure, interactive elements, etc. At the same time, the module constructs a first template demand feature vector based on the needs submitted by the user through the enterprise user interaction module, i.e., the store template demand feature information. Next, based on the template feature vector and the demand feature vector, the store template registration module applies template registration recognition multi-channel analysis. This process involves comparing the similarity of the two feature vectors and calculating the first template registration recognition coefficient, which quantifies the degree of matching between the template and the user's needs. Subsequently, it is determined whether the registration recognition coefficient reaches or exceeds the preset standard, that is, the predetermined template registration recognition coefficient. If it reaches or exceeds, this indicates that the currently traversed template is very suitable for the user's needs. Therefore, the template will be added to the first store template set, and this template set will eventually be provided to the user for final selection and further personalized adjustment.

[0050] Through the above steps, the store template registration module can efficiently screen out store templates that are highly matched with user needs, significantly improving the accuracy of template selection and user satisfaction. In addition, the automated registration and screening process greatly reduces design time and cost, and improves the efficiency of the entire template design and deployment process;

[0051] Furthermore, based on the first store template feature vector and the first template demand feature vector, the first template registration recognition coefficient is obtained according to the template registration recognition multi-channel, and the store template registration module further includes:

[0052] The template registration recognition multi-channels include a template registration breadth recognition channel, a template registration depth recognition channel and a template registration recognition combing channel.

[0053] The first store template feature vector and the first template demand feature vector are input into the template registration breadth recognition channel to obtain a first template registration breadth recognition coefficient.

[0054] Based on the first store template feature vector and the first template demand feature vector, a first template registration depth recognition coefficient is obtained according to the template registration depth recognition channel.

[0055] The first template registration breadth recognition coefficient and the first template registration depth recognition coefficient are input into the template registration recognition combing channel to generate the first template registration recognition coefficient.

[0056] Optionally, the template registration recognition multi-channel in this embodiment is composed of three main parts, namely, the template registration breadth recognition channel, the template registration depth recognition channel and the template registration recognition combing channel, each of which analyzes different feature sets to ensure a comprehensive evaluation of the degree of matching between the template and user needs. Among them, the template registration breadth recognition channel is responsible for evaluating the consistency of the overall features of the template with user needs, such as the style, color scheme and layout of the template. By inputting the first store template feature vector and the first template requirement feature vector, the channel calculates a breadth recognition coefficient, which reflects the matching degree of the template on the broad features; the template registration depth recognition channel deeply analyzes more detailed elements, such as image quality, font selection and the applicability of interactive elements. Also based on the input feature vector, the channel generates a depth recognition coefficient, indicating the degree of adaptation of the template at the detail level; and the template registration recognition combing channel is used to input the breadth and depth recognition coefficients into the combing channel, which integrates the data of the first two channels and generates the final first template registration recognition coefficient through a series of algorithm processing. This coefficient is the key indicator for determining whether to select the template to be added to the first store template set.

[0057] Through the hierarchical analysis method of template registration recognition multi-channel, the store template registration module can more accurately evaluate and select store templates that meet the needs of enterprise users. The breadth recognition channel ensures that the template meets the user's general needs as a whole, while the depth recognition channel ensures perfect adaptation in details. The final combing channel provides a comprehensive evaluation result through reasonable data fusion to ensure that the selected template is optimal on multiple levels.

[0058] Furthermore, the store template registration module also includes:

[0059] The template registration recognition combing channel includes a template registration recognition combing function, and the template registration recognition combing function is:

[0060] ;

[0061] Among them, TRC represents the template registration recognition coefficient, TRA represents the template registration breadth recognition coefficient, TRB represents the template registration depth recognition coefficient, RAW represents the registration recognition combing first predetermined weight, and RBW represents the registration recognition combing second predetermined weight.

[0062] For example, an important feature of the store template registration module is that it uses the template registration recognition combing channel to make the final evaluation and decision on the template selection. This channel uses a specific calculation formula, namely the calculation method of the template registration recognition coefficient (TRC), to ensure that the selected template most accurately reflects the user's needs. Specifically, the template registration recognition combing function is Among them, TRC represents the template registration recognition coefficient; TRA represents the template registration breadth recognition coefficient, which reflects the degree of matching between the template and user needs in breadth, such as overall style and layout; TRB represents the template registration depth recognition coefficient, which reflects the degree of detail matching of the template in depth, such as image quality and interaction details; RAW represents the first predetermined weight of registration recognition combing, and RBW represents the second predetermined weight of registration recognition combing. RAW and RBW are weight factors used to adjust the influence of breadth and depth coefficients to ensure that the importance of breadth and depth is properly balanced during the evaluation process.

[0063] For example, for an online clothing store template, TRA may be calculated based on whether the template provides sufficient visual appeal and a layout suitable for displaying clothing; TRB may focus on evaluating the template's image processing capabilities when displaying high-definition clothing pictures and the responsiveness of user interactive elements. In this model, if it is assumed that designers pay more attention to the visual appeal of the template (i.e., breadth) rather than the interactive details (i.e., depth), RAW can be assigned a higher value and RBW can be assigned a lower value. For example, RAW can be 0.7 and RBW can be 0.3, indicating that in the final TRC calculation, the influence of the breadth factor (TRA) will be greater than the depth factor (TRB).

[0064] Through this calculation formula, the template registration and recognition combing channel can provide a quantitative score that comprehensively considers breadth and depth, which helps to more comprehensively understand the degree of match between different templates and user needs. This method not only enhances the accuracy of template selection, but also makes the final selected template more able to meet the user's actual operation and visual needs.

[0065] Furthermore, based on the store product image set and the predetermined product image background set, the first store template set is rendered and optimized to generate a second store template set, and the store template optimization module is further used to:

[0066] Activate three image processing channels, wherein the three image processing channels include an image background separation channel, a predetermined background fusion channel, and an image enhancement channel.

[0067] The image background of the store product image set is removed based on the image background separation channel to obtain a first optimized store product image set.

[0068] Based on the predetermined background fusion channel, the first optimized store product image set and the predetermined product image background set are fused to generate a second optimized store product image set.

[0069] Based on the image enhancement channel, image enhancement is performed on the second optimized store product image set to obtain a third optimized store product image set.

[0070] Rendering optimization is performed on each store template in the first store template set according to the third optimized store product image set to obtain the second store template set.

[0071] Furthermore, the store template optimization module aims to improve the visual quality of store templates through a series of sophisticated image processing steps and generate a highly customized set of store templates based on the specific needs of enterprise users. This module uses three specialized image processing channels, namely image background separation channel, predetermined background fusion channel and image enhancement channel, to systematically improve product images and integrate them into store templates.

[0072] Specifically, the module first activates three core image processing channels, which work together to ensure that everything from basic image background removal to advanced image enhancement is properly processed. Among them, the image background separation channel is responsible for removing unnecessary background elements from the store product image set to make the product image more prominent. At the beginning of the operation, the channel receives the original product image set and uses advanced image segmentation technology to identify and separate the product and the background. This step is to ensure that the product remains visually consistent and attractive in any background. After background separation, the clear product image is sent to the predetermined background fusion channel, where the product image is fused with the pre-selected background set. In this process, the background selected not only matches the product style, but also enhances the overall visual effect, such as choosing a simple and elegant background for high-end jewelry. After that, the fused image enters the image enhancement channel, where various image processing techniques such as color correction, contrast adjustment and sharpening are used to further improve the image quality. This step ensures that each product image can be presented in the best state and enhance the consumer's visual experience. The processed image set, that is, the third optimized store product image set, will be used for each template in the first store template set for final rendering optimization. At this stage, each template is adjusted and rendered according to the optimized images to ensure that the visual effects of the template perfectly match the new image set. Through these fine adjustments, the final second store template set is not only more visually attractive, but also more able to meet the specific needs of the market and users.

[0073] Through the store template optimization module, the quality of product display and the visual appeal of the template are improved. This method's systematic image optimization process not only improves consumers' shopping experience, but also helps improve the brand's market competitiveness. The optimized template can better highlight the product features, which directly promotes the increase in sales conversion rate.

[0074] Furthermore, based on the second store template set, animation recommendation comparison is performed on the enterprise user to generate a user store template, and the animation recommendation comparison module is further used to:

[0075] Pairwise comparisons are performed based on the second store template set to obtain multiple template comparison degrees.

[0076] Based on the preset template comparison degree, the second store template set is compared and adjusted in combination with the multiple template comparison degrees to obtain a third store template set with a comparison degree greater than / equal to the preset template comparison degree.

[0077] Based on the enterprise user performing animation adjustment on the third store template set, a fourth store template set is obtained.

[0078] Preview and recommend the enterprise users according to the fourth store template set to obtain preliminary store templates and user feedback data.

[0079] According to the user feedback data, the preliminary store template is adjusted to obtain the user store template.

[0080] Optionally, the animation recommendation comparison module is used to design, optimize and personalize the animation effects of the store templates of enterprise users. The implementation of this module ensures that every step from the initial template selection to the final template customization is accurate and driven by user needs. The animation recommendation comparison module first starts with the second store template set, which has been image optimized. After that, the module performs a pairwise comparison operation to evaluate the similarity and difference of each pair of templates and generates multiple template comparisons, which are quantified by preset scoring criteria to ensure that the selection of each template is based on certain performance criteria. Next, based on these comparisons and the preset template comparison criteria, the module selects those templates whose comparisons are greater than or equal to the preset criteria to form a third store template set. This step ensures that all templates entering the next stage meet the initial design and functional requirements. Subsequently, the module guides the enterprise user to make detailed animation adjustments to the third store template set to further customize the visual and functional performance of the template, including the selection of animation type, speed and interactive characteristics, so that each template is more in line with the brand image and market strategy of the enterprise. After adjustment, the generated fourth store template set is used to preview and recommend to enterprise users. At this stage, users can actually view the performance of each template and provide feedback. This preview and feedback process is interactive, allowing users to evaluate the suitability of the template from a visual and operational level. Finally, based on the user's feedback data, the animation recommendation comparison module makes final adjustments to improve the preliminary store template. These adjustments may involve further animation modifications, color adjustments, or layout optimization to ensure that the final template fully meets the user's expectations and needs. Among them, the preliminary store template refers to the template selected by the corporate user. After completing these adjustments, the final user store template is confirmed, providing corporate users with a fully customized and optimized store display solution.

[0081] Through the systematic operation of the animation recommendation comparison module, not only the attractiveness and functionality of the store templates are significantly improved, but also the user-driven and feedback-sensitive nature of the template design process is ensured. In the end, the store templates obtained by corporate users are not only visually attractive, but also fully meet their business needs and market positioning in terms of functionality.

[0082] Furthermore, based on the enterprise user performing animation adjustment on the third store template set to obtain a fourth store template set, the animation recommendation comparison module is further used to:

[0083] Collect the basic information of the enterprise users and build enterprise user portraits.

[0084] Activate a pre-built prediction model of user animation preferences.

[0085] Based on the enterprise user portrait and according to the user animation preference prediction model, a user animation preference prediction result is obtained.

[0086] The third store template set is animatedly adjusted according to the user animation preference prediction result to generate the fourth store template set.

[0087] Specifically, the animation recommendation comparison module also uses the user animation preference prediction model to optimize and personalize the animation effects of store templates. This process ensures that the template not only conforms to the user's taste visually, but also meets the user's operating habits and preferences in terms of interactivity.

[0088] First, the module collects basic information about enterprise users, such as industry type, brand characteristics, historical user behavior data, and previous template usage. This information is the key data point for building enterprise user portraits. User portraits include multi-dimensional information such as the company's market positioning, target customer groups, and brand style. Subsequently, the pre-built user animation preference prediction model is activated. This model is developed based on machine learning technology and can predict users' preferences for animation style, speed, and type based on user portraits and historical preference data. During the model training process, a large amount of user feedback data and behavior analysis results are used, and algorithm optimization is used to ensure that user preferences can be accurately predicted. Finally, based on user portraits and animation preference prediction results, the module makes personalized animation adjustments to the third store template set. These adjustments include changing the type of animation (such as switching from simple sliding to complex 3D effects), adjusting the display speed of the animation, and optimizing the interactive mode of the animation. The adjusted template set is used as the fourth store template set, which will be more in line with the actual needs and preferences of users.

[0089] Furthermore, the construction process of the user animation preference prediction model includes data collection, model selection, training, testing, and iterative optimization of the model. Specifically, first, it is necessary to collect data for training and testing the model. These data usually include user interaction data, such as user click-through rate, page dwell time, interaction frequency, etc. under different animation effects; user feedback, such as directly obtaining preferences for different animation effects from user surveys or feedback; user background information, including user industry categories, historical purchase behaviors, previous template usage, etc.; and technical performance data, i.e., indicators such as loading time and performance of various animation effects. After collecting the data, useful information is extracted from the raw data, such as extracting user preferences for specific animation effects from user activity logs. Further, determine which features are most useful for the prediction model, including user behavior characteristics, historical preferences, technical indicators, etc. Next, use methods such as normalization and standardization to transform the features so that they are suitable for machine learning models. Then, select a suitable machine learning model to build a prediction model. Possible model choices include decision trees, random forests and gradient boosting machines, neural networks, support vector machines, etc. When training a model, the data set needs to be divided into a training set and a test set. The training set is used to train the model, and then the test set is used to evaluate the performance of the model. Then, the model is evaluated and optimized. The performance of the model usually depends on statistical indicators such as accuracy, recall, and F1 score. Based on the results of the test set, it may be necessary to adjust the model parameters or reselect the model. In addition, the generalization ability of the model can be more accurately evaluated using techniques such as cross-validation. Finally, according to the performance of the model in actual applications, the model is continuously iterated and optimized, which may involve retraining the model to adapt to new user data, or adding new data sources and features to the model to improve the prediction effect. The construction of the user animation preference prediction model is completed through the above steps. Building a user animation preference prediction model is a dynamic and iterative process that requires comprehensive consideration of the quality of the data, the applicability of the model, and the actual needs of the business. Through continuous optimization and adjustment, the prediction accuracy and practicality of the model can be gradually improved.

[0090] By introducing the user animation preference prediction model, the user customization level of the template has been greatly improved, so that the final store template is not only visually attractive to users, but also can achieve higher satisfaction in user experience. This prediction and personalized adjustment process helps to improve user stickiness and the conversion rate of the platform, because the animation effect of the template can accurately reflect the user's brand personality and market positioning.

[0091] Through the technical solutions of the above embodiments, the present application provides a natural language-based enterprise user template production tool, which solves the technical problems of poor market response immediacy of user templates, low design accuracy and slow response speed, and inability to dynamically adjust template personalization in real time, resulting in low user output efficiency and high costs. It achieves the goal of improving the degree of automation and personalization accuracy of template generation, while increasing the dynamic adjustment capability of templates, so that templates can reflect the latest status of store operations and market changes in real time, achieving the technical effect of improving efficiency while reducing market investment costs.

[0092] Embodiment 2 is based on the same inventive concept as the natural language-based enterprise user template production tool in the above embodiment. Figure 2 As shown, the embodiment of the present application provides a method for producing enterprise user templates based on natural language, the method comprising:

[0093] Receive template generation requirement information of an enterprise user, wherein the template generation requirement information includes a store template design requirement, a store product image set, and a predetermined product image background set of the enterprise user.

[0094] The store template design requirements are analyzed based on the NLP model to determine the store template requirement feature information.

[0095] Based on the template registration recognition multi-channel, template registration is performed on the store template demand feature information to construct a first store template set.

[0096] The first store template set is rendered and optimized based on the store product image set and the predetermined product image background set to generate a second store template set.

[0097] Based on the second store template set, animation recommendation comparison is performed on the enterprise user to generate a user store template.

[0098] The real-time store monitoring data of the enterprise user is obtained, and the user store template is adjusted in association with the real-time store monitoring data.

[0099] Further, template registration is performed on the store template demand feature information based on template registration recognition multiple channels to construct a first store template set, including:

[0100] Activate the online store template library in the store template registration module.

[0101] The online store template library is traversed to extract a first online store template.

[0102] A first shop template feature vector is constructed according to the first online shop template.

[0103] A first template demand feature vector is constructed according to the store template demand feature information.

[0104] Based on the first store template feature vector and the first template demand feature vector, a first template registration recognition coefficient is obtained according to the template registration recognition multi-channel.

[0105] It is determined whether the first template registration recognition coefficient is greater than / equal to a predetermined template registration recognition coefficient.

[0106] If the first template registration recognition coefficient is greater than / equal to the predetermined template registration recognition coefficient, the first online store template is added to the first store template set.

[0107] Further, based on the first store template feature vector and the first template demand feature vector, a first template registration recognition coefficient is obtained according to the template registration recognition multi-channel, including:

[0108] The template registration recognition multi-channels include a template registration breadth recognition channel, a template registration depth recognition channel and a template registration recognition combing channel.

[0109] The first store template feature vector and the first template demand feature vector are input into the template registration breadth recognition channel to obtain a first template registration breadth recognition coefficient.

[0110] Based on the first store template feature vector and the first template demand feature vector, a first template registration depth recognition coefficient is obtained according to the template registration depth recognition channel.

[0111] Inputting the first template registration breadth recognition coefficient and the first template registration depth recognition coefficient into the template registration recognition combing channel to generate the first template registration recognition coefficient;

[0112] Furthermore, the template registration recognition combing channel includes a template registration recognition combing function, and the template registration recognition combing function is:

[0113] ;

[0114] Among them, TRC represents the template registration recognition coefficient, TRA represents the template registration breadth recognition coefficient, TRB represents the template registration depth recognition coefficient, RAW represents the registration recognition combing first predetermined weight, and RBW represents the registration recognition combing second predetermined weight.

[0115] Further, based on the store product image set and the predetermined product image background set, the first store template set is rendered and optimized to generate a second store template set, including:

[0116] Activate three image processing channels, wherein the three image processing channels include an image background separation channel, a predetermined background fusion channel, and an image enhancement channel.

[0117] The image background of the store product image set is removed based on the image background separation channel to obtain a first optimized store product image set.

[0118] Based on the predetermined background fusion channel, the first optimized store product image set and the predetermined product image background set are fused to generate a second optimized store product image set.

[0119] Based on the image enhancement channel, image enhancement is performed on the second optimized store product image set to obtain a third optimized store product image set.

[0120] Rendering optimization is performed on each store template in the first store template set according to the third optimized store product image set to obtain the second store template set.

[0121] Further, performing animation recommendation comparison on the enterprise user based on the second store template set to generate a user store template includes:

[0122] Pairwise comparisons are performed based on the second store template set to obtain multiple template comparison degrees.

[0123] Based on the preset template comparison degree, the second store template set is compared and adjusted in combination with the multiple template comparison degrees to obtain a third store template set with a comparison degree greater than / equal to the preset template comparison degree.

[0124] Based on the enterprise user performing animation adjustment on the third store template set, a fourth store template set is obtained.

[0125] Preview and recommend the enterprise users according to the fourth store template set to obtain preliminary store templates and user feedback data.

[0126] According to the user feedback data, the preliminary store template is adjusted to obtain the user store template.

[0127] Further, based on the enterprise user performing animation adjustment on the third store template set, a fourth store template set is obtained, including:

[0128] Collect the basic information of the enterprise users and build enterprise user portraits.

[0129] Activate a pre-built prediction model of user animation preferences.

[0130] Based on the enterprise user portrait and according to the user animation preference prediction model, a user animation preference prediction result is obtained.

[0131] The third store template set is animatedly adjusted according to the user animation preference prediction result to generate the fourth store template set.

[0132] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0133] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0134] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A natural language-based enterprise user template production tool, characterized in that: The tools include: An enterprise user interaction module, the enterprise user interaction module is used to receive template generation requirement information of an enterprise user, wherein the template generation requirement information includes a store template design requirement, a store product image set and a predetermined product image background set of the enterprise user; A natural language parsing module, which is used to parse the store template design requirements based on an NLP model to determine store template requirement feature information; A store template registration module, the store template registration module is used to perform template registration on the store template demand feature information based on template registration recognition multi-channel to construct a first store template set; A store template optimization module, the store template optimization module is used to perform rendering optimization on the first store template set based on the store product image set and the predetermined product image background set to generate a second store template set; An animation recommendation comparison module, the animation recommendation comparison module is used to perform animation recommendation comparison on the enterprise user based on the second store template set to generate a user store template; A template dynamic adjustment module, the template dynamic adjustment module is used to obtain the real-time store monitoring data of the enterprise user, and make association adjustments to the user store template according to the real-time store monitoring data; Wherein, the store template registration module is also used for: Activating the online store template library in the store template registration module; Traversing the online store template library, extracting a first online store template; Constructing a first store template feature vector according to the first online store template; Constructing a first template demand feature vector according to the store template demand feature information; Based on the first store template feature vector and the first template demand feature vector, obtaining a first template registration recognition coefficient according to the template registration recognition multi-channel; Determining whether the first template registration recognition coefficient is greater than / equal to a predetermined template registration recognition coefficient; If the first template registration recognition coefficient is greater than / equal to the predetermined template registration recognition coefficient, adding the first online store template to the first store template set; The template registration recognition multi-channel includes a template registration breadth recognition channel, a template registration depth recognition channel and a template registration recognition combing channel; Inputting the first store template feature vector and the first template demand feature vector into the template registration breadth recognition channel to obtain a first template registration breadth recognition coefficient; Based on the first store template feature vector and the first template demand feature vector, and according to the template registration depth recognition channel, obtaining a first template registration depth recognition coefficient; The first template registration breadth recognition coefficient and the first template registration depth recognition coefficient are input into the template registration recognition combing channel to generate the first template registration recognition coefficient.

2. A natural language-based enterprise user template production tool as claimed in claim 1, characterized in that: The store template registration module also includes: The template registration recognition combing channel includes a template registration recognition combing function, and the template registration recognition combing function is: ； Among them, TRC represents the template registration recognition coefficient, TRA represents the template registration breadth recognition coefficient, TRB represents the template registration depth recognition coefficient, RAW represents the registration recognition combing first predetermined weight, and RBW represents the registration recognition combing second predetermined weight.

3. The natural language-based enterprise user template production tool according to claim 1, characterized in that: The store template optimization module is also used for: Activating three image processing channels, wherein the three image processing channels include an image background separation channel, a predetermined background fusion channel, and an image enhancement channel; Based on the image background separation channel, the image background of the store product image set is removed to obtain a first optimized store product image set; Based on the predetermined background fusion channel, performing image fusion on the first optimized store product image set and the predetermined product image background set to generate a second optimized store product image set; Based on the image enhancement channel, performing image enhancement on the second optimized store product image set to obtain a third optimized store product image set; Rendering optimization is performed on each store template in the first store template set according to the third optimized store product image set to obtain the second store template set.

4. The natural language-based enterprise user template production tool according to claim 1, characterized in that: The animation recommendation comparison module is also used for: Perform pairwise comparisons according to the second store template set to obtain multiple template comparison degrees; Based on a preset template comparison degree, the second store template set is compared and adjusted in combination with the multiple template comparison degrees to obtain a third store template set having a comparison degree greater than / equal to the preset template comparison degree; Based on the enterprise user performing animation adjustment on the third store template set, a fourth store template set is obtained; Preview and recommend the enterprise user according to the fourth store template set to obtain a preliminary store template and user feedback data; According to the user feedback data, the preliminary store template is adjusted to obtain the user store template.

5. A natural language-based enterprise user template production tool as claimed in claim 4, characterized in that: The animation recommendation comparison module is also used for: Collect basic information of the enterprise users and build enterprise user portraits; Activate a pre-built prediction model of user animation preferences; Based on the enterprise user portrait and the user animation preference prediction model, obtaining a user animation preference prediction result; The third store template set is animatedly adjusted according to the user animation preference prediction result to generate the fourth store template set.

6. A method for producing enterprise user templates based on natural language, characterized in that: The method is applied to a natural language-based enterprise user template production tool as described in any one of claims 1 to 5, and the method comprises: Receiving template generation requirement information from an enterprise user, wherein the template generation requirement information includes a store template design requirement, a store product image set, and a predetermined product image background set from the enterprise user; Analyze the store template design requirements based on the NLP model to determine the store template requirement feature information; Performing template registration on the store template demand feature information based on template registration recognition multiple channels to construct a first store template set; Performing rendering optimization on the first store template set based on the store product image set and the predetermined product image background set to generate a second store template set; Performing animation recommendation comparison for the enterprise user based on the second store template set to generate a user store template; The real-time store monitoring data of the enterprise user is obtained, and the user store template is adjusted in association with the real-time store monitoring data.

Images