A customer-oriented customized decoration interaction method and system

Abstract

This application relates to a customized home decoration interaction method and system for customers. By collecting customer online browsing behavior data and extracting key features such as click frequency and dwell time, a preset threshold is used to determine whether an "ignition" condition is triggered. When the condition is met, the system generates ignition information and updates the ignition rule base in real time based on customer interest levels and historical data comparison, thereby accurately grasping the changing trends of customer preferences. Based on personalized solutions, the system constructs virtual reality scenes, generates 3D display models, and ensures consistent model display across different devices through scene optimization algorithms and terminal device adaptation technology. Finally, a loading time minimization strategy is adopted to achieve fast loading and resource conservation. This invention effectively solves the problems of slow response and low personalization in traditional home decoration display systems, significantly improving customer experience and system efficiency.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a customized decoration interaction method and system for customers. Background Technology

[0002] In existing technologies, customized home decoration is designed and manufactured based on consumers' individuality and preferences, which can meet most of their personalized needs and match the apartment layout well in terms of size, style, design, and function. However, the production and manufacturing of customized home decoration is affected by the human resources in the design stage, resulting in low production efficiency and difficulty in meeting market demand. In recent years, the rapid development of Internet technology and the improvement of people's consumption level have led more and more people to choose to obtain and browse information online. The e-commerce of the home decoration market is becoming more and more comprehensive. Recommendation engines are similar to product recommenders in our real lives, providing great convenience for people to choose suitable home decoration plans during the home decoration process. Traditional personalized home decoration recommendation algorithms mostly rely on the explicit interaction records of users to obtain recommendation results, which often leads to a lack of timeliness and freshness in the recommendation results. Furthermore, the diverse and constantly changing interests of users make it difficult to accurately represent them, and the insufficient mining of semantic relationships affects the accuracy and generalization ability of the recommendation model. Therefore, it is necessary to deeply understand user behavior and specific actual situations to provide home decoration plan recommendation services for users with decoration needs. How to accurately express user interests and improve the accuracy and targeting of recommendations during the home decoration plan recommendation process is an urgent problem to be solved.

[0003] In customized home decoration services, "ignition" is a key concept, referring to the mechanism that triggers personalized recommendations based on customer behavior. However, achieving accurate "ignition" and subsequent 3D display faces a series of technical challenges. The primary issue is how to quickly identify effective "ignition" features from massive amounts of customer data and update the rule base in real time to adapt to dynamic changes in customer preferences. This requires striking a balance between data collection, feature extraction, and model updates, ensuring both timeliness and accuracy in recommendations. Secondly, the generation and optimization of personalized 3D display models also present contradictions. On the one hand, models are required to have high realism and rich detail; on the other hand, the hardware limitations and network conditions of different terminal devices must be considered. How to reduce model complexity while ensuring visual quality and achieving consistent display across platforms is a thorny problem. Furthermore, real-time rendering and loading of 3D models also face challenges, requiring minimizing resource consumption and network transmission overhead while ensuring a smooth user experience. The core of these problems lies in how to achieve a highly personalized, high-quality 3D visualization experience with limited computing and network resources, while ensuring the system's real-time responsiveness and cross-platform compatibility. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, this invention provides a customized decoration interaction method and system for customers.

[0005] The technical solution adopted by this invention to solve its technical problem is:

[0006] This invention provides a customized interior design interaction method for customers, mainly including:

[0007] The system acquires initial selection and browsing behavior data of customers within the system, extracting key features from this data, including click frequency, dwell time, and page navigation paths. For these extracted key features, preset thresholds are used to determine if the customer behavior meets the trigger conditions. If it does, trigger conditions are generated; otherwise, data collection continues. Based on the trigger conditions, the system assesses the customer's level of interest in specific content or functions within the customized renovation, comparing this interest with historical data in the trigger rule base to determine trends in customer preferences. The trigger rule base is updated in real-time. Finally, a personalized renovation plan is generated based on the trigger rule base, constructing a virtual reality scene. The 3D models in the scene are rendered using a pre-defined material texture library to generate an initial 3D display model. For this initial 3D display model, a scene construction optimization algorithm is used to simplify the model, reducing the number of polygons and material complexity, resulting in an optimized 3D display model. Based on the optimized 3D display model, a terminal device adaptation algorithm is used to evaluate the hardware performance of different devices, adjusting the model resolution and rendering parameters to ensure consistent display effects. After terminal device adaptation is complete, a loading time minimization strategy is used to compress and segment the 3D display model, reducing network transmission time and resource consumption, resulting in the final display model.

[0008] This invention provides a customized home renovation interaction system for customers, mainly comprising: a data acquisition module, used to acquire initial selection and browsing behavior data of customers in the system, and extract key features from the customer behavior data, including click frequency, dwell time, and page jump path; a feature extraction module, used to determine whether the customer behavior meets the trigger ignition conditions based on the extracted key features using preset thresholds; if it meets the conditions, trigger ignition information is generated; otherwise, data acquisition continues; an ignition judgment module, used to determine the customer's level of interest in specific content or functions in the customized renovation based on the trigger ignition information, and compare it with historical data in the ignition rule base to determine the trend of customer preference changes, while updating the ignition rule base in real time; and a 3D model construction module, used to construct individual models based on the generated features. A personalized decoration scheme is used to construct a virtual reality scene. A preset material texture library is used to render the 3D models in the scene, generating an initial 3D display model. A 3D model optimization module simplifies the initial 3D display model using a scene construction optimization algorithm, reducing the number of polygons and material complexity to generate an optimized 3D display model. A device adaptation module evaluates the hardware performance of different devices using a terminal device adaptation algorithm based on the optimized 3D display model, adjusting the model resolution and rendering parameters to ensure consistent display effects. A model distribution optimization module compresses and segments the 3D display model using a loading time minimization strategy after terminal device adaptation is complete, reducing network transmission time and resource consumption to obtain the final display model. The technical solution provided by this invention can include the following beneficial effects:

[0009] This invention discloses a customized home decoration interaction method for customers. By collecting online browsing behavior data from customers and extracting key features such as click frequency and dwell time, a preset threshold is used to determine whether an "ignition" condition is triggered. When the condition is met, the system generates ignition information and updates the ignition rule base in real time based on customer interest levels and historical data comparison, thereby accurately grasping the changing trends of customer preferences. Based on personalized solutions, the system constructs virtual reality scenes, generates 3D display models, and ensures consistent display of the models on different devices through scene optimization algorithms and terminal device adaptation technology. Finally, a loading time minimization strategy is adopted to achieve fast loading and resource conservation. This invention effectively solves the problems of slow response and low personalization in traditional home decoration display systems, significantly improving customer experience and system efficiency. Attached Figure Description

[0010] Figure 1 This is a flowchart of a customer-oriented customized decoration interaction method according to the present invention.

[0011] Figure 2 This is a schematic diagram of a customer-oriented customized decoration interaction method and system according to the present invention.

[0012] Figure 3 This is another schematic diagram of a customer-oriented customized decoration interaction method and system according to the present invention.

[0013] Figure 4 This is a schematic diagram of a customer-oriented customized decoration interaction method and system according to the present invention.

[0014] Figure 5 This is an example of one presentation format of the front-end interface of the customized decoration interactive system designed for this invention.

[0015] Figure 6 This is an example of another presentation form of the front-end interface of the customized decoration interaction system designed in this invention.

[0016] Figure 7 Example of a front-end interface diagram for a customized decoration interaction system designed for this invention, allowing users to browse the effects of building materials of different materials in real time.

[0017] Figure 8 This invention provides a customized interior design interactive system that allows users to view 3D renderings of different interior design effects in real time.

[0018] Figure 9 This is a schematic diagram of the backend database module design for the customized decoration interaction system designed for this invention. Detailed Implementation

[0019] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1-4 This embodiment of a customized decoration interaction method and system for customers may specifically include:

[0021] Step S101: Obtain the customer's initial selection and browsing behavior data in the system, and extract key features from the customer behavior data, including click frequency, dwell time, and page jump path.

[0022] Acquire initial selection and browsing behavior data of customers within the system, extracting click frequency, dwell time, and page navigation paths as key features. Based on click frequency and dwell time, clustering algorithms are used to classify customer behavior, resulting in grouped customer behavior. For page navigation paths, association rule mining algorithms are used to analyze frequent patterns in customer browsing paths, determining the regularity of customer browsing behavior. Combining the initial selection data, decision tree algorithms are used to predict customer behavior, determining the pages or functions customers are likely to select. Based on the clustering results and association rules, customer behavior profiles are generated, describing typical customer behavior characteristics within the system. Using the decision tree prediction results and behavior profiles, a customer behavior recommendation model is constructed to optimize page layout and function recommendations. The recommendation model is combined with business data to generate a customer behavior analysis report for subsequent business decision support.

[0023] For example, by embedding event tracking code in the user's system—such as deploying JavaScript code on key events like page load, button clicks, and page scrolling—you can capture the user's initial selections and browsing behavior and begin tracking their browsing behavior on the product page. (See attached...) Figure 8 The image illustrates a product preview page where users can browse different renderings and navigate between them. For example, a user might spend 60 seconds on rendering page A, browsing details of 3 home renovation products, then navigate to rendering page B, spend 100 seconds browsing details of 5 home renovation products, and finally return to the homepage. This entire sequence of actions is meticulously recorded. Next, feature extraction is performed on this data. Click frequency is calculated by counting the number of clicks per unit of time. For example, if a user clicks 15 times in 10 minutes, the click frequency is 5 clicks per minute. Comparing this with the behavior of users of similar age, identity, and background provides a more comprehensive understanding of user behavior. Dwell time records the time a user spends on each page, such as the 60 seconds and 100 seconds mentioned above. The page navigation path is represented by a directed graph, with nodes representing pages and edges representing navigation actions. For example, "Homepage" -> "Rendering A Page" -> "Rendering B Page" -> "Homepage" constitutes a path.

[0024] Step S102: For the extracted key features, a preset threshold is used to determine whether the customer's behavior meets the conditions for triggering ignition. If it does, triggering ignition information is generated; otherwise, data collection continues.

[0025] Acquire customer behavior data and extract key features from it. Determine if the key features meet trigger conditions based on preset thresholds. If the key features meet the trigger conditions, generate ignition information. If the key features do not meet the trigger conditions, continue collecting customer behavior data. Update the key feature extraction rules using behavioral analysis methods. Dynamically adjust preset thresholds through a threshold comparison module. Based on the condition matching results, determine whether to generate ignition information or continue data collection.

[0026] Based on customer behavior data, key feature extraction rules are used to obtain the attributes related to the triggering conditions. The threshold comparison module is used to determine whether the ignition information generation standard is met. If it is not met, data collection continues and the preset threshold is dynamically adjusted until the condition matching result is determined and the ignition information is generated.

[0027] A customer behavior data acquisition module acquires customer behavior data, which is then input into a key feature extraction module to extract trigger condition-related attributes. The extracted attributes are then input into a threshold comparison module to determine if they meet the ignition information generation criteria. If they do, ignition information is generated; otherwise, the data acquisition module continues to collect data, and a dynamic adjustment module is invoked to update the preset threshold. Based on the updated preset threshold, threshold comparisons are performed again until a matching result is obtained, determining whether ignition information is generated. Machine learning algorithms are used to update the key feature extraction rules, optimizing the extraction accuracy of trigger condition-related attributes. Dynamic adjustment of the preset threshold improves the accuracy and timeliness of ignition information generation.

[0028] The above steps can be achieved through, for example, appendix Figure 7 The illustration shows multiple sets of parallel building material effect selection pages, guiding users to repeatedly switch and compare, and recording browsing and UI interaction data for analysis. Note the appendix. Figure 7 The illustrated front-end page effect is not the only presentation method limited by this invention. Step S103: Based on the trigger ignition information, determine the customer's level of interest in specific content or functions in the customized decoration, compare it with historical data in the ignition rule base to determine the trend of customer preference changes, and update the ignition rule base in real time.

[0029] The system acquires customer ignition information and extracts behavioral data related to specific content or functions available in customized decoration. Based on the extracted behavioral data, an interest-based calculation model is used to quantify the customer's level of interest in specific content or functions. Historical data is retrieved from the ignition rule base, and the current level of interest is compared and analyzed with historical data. If there is a significant difference between the current level of interest and historical data, it is determined that the customer's preference has changed. Based on the trend of preference changes, a dynamic rule update algorithm is used to adjust the rules in the ignition rule base in real time. Through data comparison and analysis, changes in customer behavior patterns are identified, and new preference tags are generated. The new preference tags are matched with the updated rule base to generate personalized recommendations for the customer.

[0030] For example, obtaining customer trigger information is a key step in understanding user needs. In a customized renovation scenario, this might include browsing images of home furnishings in a specific style, checking prices for certain materials, or saving specific design schemes. For instance, the system might record that a user frequently browses Nordic-style living room designs and repeatedly views information related to solid wood flooring. After extracting behavioral data, quantifying customer preferences using an interest-based calculation model is crucial. This model might consider multiple factors, such as browsing duration, interaction frequency, and search keywords. Assuming the system assigns weights to different behaviors, such as giving a higher weight to saving design schemes than simply browsing, through comprehensive analysis, the system might determine that the user's interest in Nordic style is 8.5 (out of 10) and their interest in solid wood flooring is 7.2. Comparing current interest levels with historical data is an effective way to identify changes in preferences. The trigger rule base might store user interest data from the past three months. If it's found that a user's interest in Nordic style jumps from 5.6 to 8.5, while their interest in traditional Chinese style drops from 7.8 to 3.2, this indicates a significant change in user preferences. Dynamically updating the rule base based on changing preference trends is key to maintaining the sensitivity of a recommendation system. For example, the system might use a sliding window approach, giving higher weight to data from the most recent week while gradually reducing the influence of older data. This way, if a user has frequently viewed solid wood flooring in the past week, even if they previously showed a high interest in carpets, the system will quickly adjust its recommendation strategy. Identifying changes in customer behavior patterns and generating new preference tags is fundamental to personalized recommendations. For instance, the system might discover that a user is not only interested in Scandinavian style but also pays particular attention to its minimalist elements. Therefore, in addition to the existing "Skander style enthusiast" tag, a new "minimalist pursuer" tag might be added. Matching the new preference tags with the updated rule base generates more accurate personalized recommendations. For example, the system might recommend furniture in line with Scandinavian minimalism, such as a white sofa with clean lines and a simple, natural wood dining table. Simultaneously, considering the user's interest in solid wood flooring, light oak flooring might also be recommended as a complement. This dynamically adjusted recommendation mechanism can promptly capture subtle changes in user preferences, providing more personalized services tailored to their needs. It not only improves user experience but also helps businesses better understand market trends, providing valuable insights for product development and inventory management. By continuously optimizing algorithms and expanding data dimensions, this system can create a significant advantage for businesses in the highly competitive customized decoration market.

[0031] Based on customer behavior data, after generating trigger ignition information, a multimodal big data model is used to comprehensively analyze customer behavior and feedback information. This multimodal big data model can extract deep semantic features from the customer's facial expressions, posture, voice tone, natural language input, and existing browsing behavior data. Combined with historical data in the ignition rule base, it judges the customer's interest in specific content or functions, determines the trend of customer preference changes, and updates the ignition rule base in real time. An interest degree calculation model is used to judge the trend of customer preference changes, and the rule base is adjusted in real time through a dynamic rule update algorithm to obtain new personalized recommendation content. Personalized decoration plans are generated through personalized recommendation content in the rule base.

[0032] Step S104: Based on the generated personalized decoration plan, construct a virtual reality scene, use a preset material texture library to render the 3D model in the scene, and generate an initial 3D display model.

[0033] For example, assuming the system generates a personalized decoration scheme including elements such as "modern minimalist style," "intelligent lighting control," and "open kitchen" based on the aforementioned steps, the next step is to construct a corresponding virtual reality scene. First, the system calls a pre-established 3D model library containing 3D models of various furniture, building materials, and appliances, such as the sofa model numbered "SF-00123," the smart light bulb model numbered "LT-00456," and the cabinet model numbered "CB-00789." Based on elements in the scheme, such as the "L-shaped gray fabric sofa," the "Philips Hue smart light bulb," and the "white lacquered cabinet," the system automatically matches the corresponding 3D models and loads them into the virtual scene, arranging them according to the spatial layout specified in the scheme. For example, the sofa is located in the center of the living room, the smart light bulb is installed on the ceilings of the living room and bedroom, and the cabinets are arranged in an L-shape around the kitchen walls. Then, the system calls a material mapping library containing high-resolution texture images of various materials and their corresponding physical property parameters, such as roughness, metallicity, and normal maps. For example, material number "MTL-001" represents "matte gray fabric," with a roughness parameter set to 8 and a metallicity parameter set to 1, corresponding to a 512x512 pixel gray fabric texture image; material number "MTL-002" represents "white paint," with a roughness parameter set to 2 and a metallicity parameter set to 9, corresponding to a 512x512 pixel white highlight texture image. According to the description in the solution, the system assigns material "MTL-001" to the "SF-00123" sofa model and material "MTL-002" to the "CB-00789" cabinet model. To facilitate viewing on different devices, the system also converts the 3D display models to the universal glTF format and compresses them to approximately 30% of their original size using the Draco algorithm; for example, a model originally 10MB is compressed to 3MB, facilitating fast loading and smooth display. Simultaneously, to ensure data security, the system performs AES-256 encryption on the 3D model files.

[0034] As attached Figure 5 As shown, a virtual reality scene is constructed, the initial 3D display model is imported into the scene, and a virtual reality decoration plan that customers can interact with is generated.

[0035] Step S105: For the initial 3D display model, a scene construction optimization algorithm is used to simplify the model, reduce the number of polygons and material complexity, and generate an optimized 3D display model.

[0036] A scene construction optimization algorithm is used to simplify the initial 3D display model, reducing the number of polygons and material complexity to generate an optimized 3D display model. Based on the optimized model, a rendering efficiency evaluation algorithm is used to calculate the model's rendering efficiency. If the rendering efficiency is lower than a preset threshold, a computing resource allocation algorithm is used to reallocate computing resources. A visual quality evaluation algorithm is used to determine whether the visual quality of the optimized model meets the requirements. If the visual quality does not meet the requirements, a material complexity adjustment algorithm is used to further optimize the materials. Based on the optimized material complexity, a polygon count adjustment algorithm is used to dynamically adjust the number of polygons in the model. Finally, an optimized 3D display model that meets both rendering efficiency and visual quality requirements is generated.

[0037] For example, scene building optimization algorithms improve rendering efficiency by simplifying the geometry and materials of 3D models. For instance, a sofa model containing 100,000 polygons can be reduced to 20,000 polygons using a quadrilateral simplification algorithm, while maintaining an appearance similarity of over 95%.

[0038] Material simplification can reduce the resolution of a 4K texture to 1K and merge multiple material channels, such as combining diffuse, normal, and specular maps into a single PBR map. Rendering efficiency evaluation algorithms calculate the rendering time per frame. Assuming the goal is to achieve 60 FPS on a mid-range device, the rendering time per frame should not exceed 16.7 milliseconds. If the actual rendering time is 25 milliseconds, further optimization is needed. Computational resource allocation algorithms may adjust GPU thread allocation, distributing more resources to critical rendering stages such as geometry processing and pixel shading. Visual quality evaluation algorithms determine quality by comparing the image differences between the original and optimized models. The Structural Similarity Index (SSIM) can be used as an evaluation metric, with SSIM values ​​ranging from -1 to 1; a higher value indicates higher similarity. If the optimized model's SSIM value is below 0.9, further adjustments are required. Material complexity adjustment algorithms may employ adaptive texture compression techniques.

[0039] For example, attached Figure 6The schematically presented overall model shows that for complex wood-textured floor surfaces, a higher texture resolution can be retained; while for walls with more uniform color, the texture resolution can be significantly reduced. This method reduces overall texture memory usage while maintaining key details. The polygon count adjustment algorithm dynamically adjusts the level of detail based on the model's importance and visibility in the scene. For example, a lower polygon count model version can be used for distant background buildings, while a higher polygon count is retained for nearby main furniture to ensure clear details. This LOD (Level of Detail) technology can improve overall rendering efficiency while maintaining visual quality. Through this series of optimization steps, the final generated 3D display model meets both rendering efficiency requirements and maintains good visual quality. For example, the optimized living room scene may contain a total of 500,000 polygons, using 20 1K resolution PBR material maps, and can be rendered at a stable 60FPS frame rate on mid-range devices while maintaining 95% visual similarity to the original high-precision model. This optimization not only improves the user experience but also reduces device requirements, allowing more users to smoothly experience 3D interior design display effects.

[0040] Step S106: Based on the optimized 3D display model, the hardware performance of different devices is evaluated using a terminal device adaptation algorithm, and the model resolution and rendering parameters are adjusted to ensure consistent display effects.

[0041] A terminal device adaptation algorithm is used to evaluate the optimized customized 3D decoration display model and obtain hardware performance data for different devices. Based on the hardware performance data, the model resolution and rendering parameters are adjusted to generate a 3D display model adapted to different devices. A rendering efficiency evaluation algorithm is used to calculate the rendering efficiency of the adapted model. If the rendering efficiency is lower than a preset threshold, a computing resource allocation algorithm is used to reallocate computing resources. A visual quality evaluation algorithm is used to determine whether the visual quality of the adapted model meets the requirements. If the visual quality does not meet the requirements, a material complexity adjustment algorithm is used to further optimize the materials. Based on the optimized material complexity, a polygon count adjustment algorithm is used to dynamically adjust the number of polygons in the model. Finally, an adapted 3D display model that meets the requirements of rendering efficiency and visual quality is generated.

[0042] Step S107: After the terminal device adaptation is completed, the 3D display model is compressed and segmented using a loading time minimization strategy to reduce network transmission time and resource consumption, and the final display model is obtained.

[0043] The 3D display model adapted for terminal devices is compressed using a compression algorithm to obtain compressed model data. Based on the compressed model data, a block-based processing algorithm divides the model into multiple blocks, generating block-based model data. Using the block-based model data, network transmission time and resource consumption are calculated to determine if a preset loading time threshold is met. If the loading time exceeds the preset threshold, a dynamic block adjustment algorithm is used to optimize the block size, generating optimized block data. Based on the optimized block data, network transmission time and resource consumption are recalculated to determine if the loading time requirement is met. If the requirement is met, the final display model is generated; otherwise, a material compression algorithm is used to further compress the model materials, generating optimized material data. Based on the optimized material data, the block-based processing is repeated to generate the final display model.

[0044] In addition to the above-mentioned interaction methods, this application also connects to or links to well-known basic data in the field, such as BIM systems, mobile front-end applications, PC front-end applications, and web front-end applications, and designs management back-ends and marketing material interfaces, user information dashboards, and other basic information components, which can constitute the attached... Figure 9 This illustrates the comprehensive basic IT support layer.

[0045] Reference Appendix Figure 4As another aspect of this application, this application provides a customized decoration interaction system for customers, mainly including: a data acquisition module, used to acquire customers' initial selection and browsing behavior data in the system, and extract key features from the customer behavior data, including click frequency, dwell time, and page jump path; a feature extraction module, used to determine whether the customer behavior meets the trigger ignition conditions based on the extracted key features using a preset threshold; if it meets the conditions, trigger ignition information is generated; otherwise, data acquisition continues; an ignition judgment module, used to determine the customer's level of interest in specific content or functions in customized decoration based on the trigger ignition information, and compare it with historical data in the ignition rule base to determine the trend of customer preference changes, while updating the ignition rule base in real time; and a 3D model construction module, used for... Based on the generated personalized decoration plan, a virtual reality scene is constructed. A preset material texture library is used to render the 3D models in the scene, generating an initial 3D display model. A 3D model optimization module simplifies the initial 3D display model using a scene construction optimization algorithm, reducing the number of polygons and material complexity to generate an optimized 3D display model. A device adaptation module evaluates the hardware performance of different devices using a terminal device adaptation algorithm based on the optimized 3D display model, adjusting the model resolution and rendering parameters to ensure consistent display effects. A model distribution optimization module compresses and segments the 3D display model using a loading time minimization strategy after terminal device adaptation is complete, reducing network transmission time and resource consumption to obtain the final display model.

[0046] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0047] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0048] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0049] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0050] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0051] The above description of the embodiments is only for the purpose of helping to understand the technical solutions and core ideas of this application; those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A customer-oriented customized decoration interaction method, characterized in that, The method comprises: acquiring initial selection and browsing behavior data of the customer in the system, extracting key features in the customer behavior data, the key features comprising click frequency, dwell time and page jump path; for the extracted key features, using a preset threshold to judge whether the customer behavior meets the triggering ignition condition, if yes, generating triggering ignition information, if not, continuing to collect data; acquiring the triggering ignition information of the customer, and extracting behavior data related to specific content or function that can be selected in customized decoration; using an interest degree calculation model to quantify the interest degree of the customer to the specific content or function according to the extracted behavior data; acquiring historical data from the ignition rule library, and comparing and analyzing the current interest degree with the historical data; if there is a significant difference between the current interest degree and the historical data, it is judged that the customer preference has changed; using a dynamic rule updating algorithm to adjust the rules in the ignition rule library in real time according to the preference change trend; recognizing the change of the customer behavior mode through data comparison and analysis, and generating a new preference label; matching the new preference label with the updated rule library to generate personalized recommendation content for the customer; generating a personalized decoration scheme according to the ignition rule library, constructing a virtual reality scene, rendering 3D models in the scene using a preset material map library to generate an initial 3D display model; in this step, the system calls a pre-established 3D model library, automatically matches corresponding 3D models according to the elements in the scheme and loads them into the virtual scene, and places them according to the space layout specified in the scheme; using a scene construction optimization algorithm to simplify the model to reduce the number of polygons and the complexity of the material, and generating an optimized 3D display model for the initial 3D display model; using a terminal device adaptation algorithm to evaluate the hardware performance of different devices according to the optimized 3D display model, adjusting the model resolution and rendering parameters to ensure the consistency of the display effect; this step includes: using a terminal device adaptation algorithm to evaluate the optimized customized decoration 3D display model, and obtaining hardware performance data of different devices; adjusting the model resolution and rendering parameters according to the hardware performance data to generate 3D display models adapted to different devices; after the terminal device adaptation is completed, using a loading time minimization strategy to compress and block the 3D display model to reduce network transmission time and resource consumption, and obtaining the final display model.

2. The method of claim 1, wherein, The acquisition of initial selection and browsing behavior data of the customer in the system and the extraction of key features in the customer behavior data specifically comprises: acquiring initial selection and browsing behavior data of the customer in the system, and extracting click frequency, dwell time and page jump path as key features; using a clustering algorithm to classify customer behavior according to click frequency and dwell time to obtain grouping results of customer behavior; using an association rule mining algorithm to analyze the frequent patterns of customer browsing paths to determine the regularity of customer browsing behavior for the page jump path; using a decision tree algorithm to predict customer behavior in combination with initial selection data to judge the pages or functions that the customer is likely to select; According to the clustering results and the association rules, a customer behavior portrait is generated to describe the typical behavior characteristics of the customer in the system; By combining the recommendation model with business data, a customer behavior analysis report is generated for subsequent business decision support. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated.

3. The method of claim 1, wherein, According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated. According to the customer behavior data, the key feature extraction rules are used to obtain the attributes related to the trigger condition, and the threshold comparison module is used to determine whether the point ignition information generation standard is met, if not, the data collection continues and the preset threshold is dynamically adjusted until the condition matching result determines that the point ignition information is generated.

4. The method of claim 3, wherein, ​ ​ 5. The method of claim 1, wherein, ​ ​ ​ ​ 6. The method of claim 1, wherein, ​ The scene construction optimization algorithm is used to simplify the initial 3D display model, reduce the number of polygons and the material complexity, and generate an optimized 3D display model; According to the optimized model, the rendering efficiency evaluation algorithm is used to calculate the rendering efficiency of the model; If the rendering efficiency is lower than the preset threshold, the computing resource allocation algorithm is used to re-allocate the computing resources; The visual quality evaluation algorithm is used to determine whether the visual quality of the optimized model meets the requirements; If the visual quality does not meet the requirements, the material complexity adjustment algorithm is used to further optimize the material; According to the optimized material complexity, the polygon number adjustment algorithm is used to dynamically adjust the number of polygons of the model; Finally, an optimized 3D display model that meets the requirements of rendering efficiency and visual quality is generated.

7. A customer-oriented customizing furnishing interaction system for implementing the method according to any one of claims 1 to 6, characterized in that, The system comprises a data acquisition module for acquiring initial selection and browsing behavior data of customers in the system, extracting key features in customer behavior data, the key features including click frequency, stay time and page jump path; a feature extraction module for judging whether the customer behavior meets the trigger ignition condition by using a preset threshold for the extracted key features, generating trigger ignition information if it meets the condition, or continuing to collect data if it does not meet the condition; an ignition judgment module for judging the interest degree of customers in specific content or function in customized decoration according to the trigger ignition information, comparing it with historical data in the ignition rule library, determining the customer preference change trend, and updating the ignition rule library in real time; a 3D model construction module for constructing a virtual reality scene according to the generated personalized decoration scheme, rendering the 3D model in the scene by using a preset material map library, and generating an initial 3D display model; a 3D model optimization module for simplifying the initial 3D display model by using a scene construction optimization algorithm, reducing the number of polygons and the material complexity, and generating an optimized 3D display model; a device adaptation module for evaluating the hardware performance of different devices by using a terminal device adaptation algorithm according to the optimized 3D display model, adjusting the model resolution and rendering parameters to ensure the consistency of the display effect; a model distribution optimization module for compressing and blocking the 3D display model by using a loading time minimization strategy after the terminal device adaptation is completed, reducing the network transmission time and resource consumption, and obtaining the final display model.

Images