A multimodal home space design system based on a complete home furnishing dataset.

Abstract

This invention discloses a multimodal home space design system based on a complete home furnishing dataset, belonging to the field of home space design technology. The system includes: a complete home furnishing dataset management module, used to construct and manage the complete home furnishing dataset; an intelligent floor plan processing module, used to acquire floor plan data and divide it into zones, and perform furniture-filling design on the divided floor plan based on user attribute data to achieve integrated molding; and a furniture layout generation module, used to generate baselines on the floor plan after furniture filling, and fill in furniture and appliance models based on the baselines. This invention integrates a complete home furnishing dataset with intelligent algorithms to achieve automatic floor plan recognition, intelligent scheme generation, virtual visualization, and internal supply chain collaboration, solving the problems of data silos, low efficiency, and insufficient user experience in traditional design, and improving the intelligence level of home design and customer satisfaction.

Description

Technical Field

[0001] This invention relates to the field of home space design technology, specifically a multimodal home space design system based on a complete home furnishing dataset. Background Technology

[0002] With the advancement of digitalization in the home decoration industry, home space design is gradually moving towards intelligence and personalization, placing increasing demands on efficiency in floor plan recognition, solution generation, and supply chain collaboration. The industry generally relies on multimodal data (such as floor plans, design style parameters, and information on main and auxiliary materials) to support the design process, while simultaneously meeting users' needs for rapid design solution generation, personalized adaptation, and intuitive previews. This makes building a design system that integrates data resources and intelligent algorithms a trend.

[0003] Currently, existing home design technologies suffer from data fragmentation. Data on apartment layouts, styles, and main and auxiliary materials are scattered across different systems, lacking a standardized integration mechanism. This results in low data retrieval efficiency during the design process, making it difficult to support the generation of intelligent solutions. Furthermore, the design process relies excessively on manual operation. Apartment layout identification, zoning, and furniture arrangement must be completed manually, which is not only time-consuming but also leads to low matching between the design and user attributes (such as family composition) and supply chain resources. This can result in poor design feasibility and insufficient user satisfaction. Summary of the Invention

[0004] The purpose of this invention is to provide a multimodal home space design system based on a complete home furnishing dataset to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a multimodal home space design system based on a complete home furnishing dataset, comprising:

[0006] The integrated home furnishing dataset management module is used to build and manage integrated home furnishing datasets, which are structured databases containing apartment layout data, design style data, main and auxiliary material data, and home furnishing product matching data.

[0007] The intelligent apartment layout processing module is used to acquire apartment layout data and divide it into zones. Based on user attribute data, it performs home furnishing design on the divided apartment layout to achieve integrated molding.

[0008] The furniture layout generation module is used to generate baselines on the floor plan after furniture filling, and to fill in furniture and appliance models based on the baselines;

[0009] The virtual visualization and interaction module is used to display design schemes in the form of 3D models and supports interactive operations on the design schemes.

[0010] The internal supply chain collaboration module is used to enable data sharing and process collaboration between the design system and the internal supply chain management system.

[0011] Preferably, the intelligent apartment layout processing module includes:

[0012] The apartment layout zoning unit is used to divide the apartment data into zones using a spatial semantic segmentation algorithm. The zoning is based on lighting conditions, circulation distance, and functional attributes to determine the zoning boundaries.

[0013] The user attribute parsing unit is used to receive user attribute data, which includes family member composition and usage needs.

[0014] The home furnishing filling unit is used to fill the floor, walls and ceiling of the divided floor plan according to the user attribute data and the whole-house furnishing dataset, so that the space layout is full and the functions are complete.

[0015] Preferably, the zoning rules for the apartment layout include: dividing areas with a daylight factor greater than or equal to a preset value into public activity areas, dividing areas with a circulation distance less than or equal to a preset value into core functional areas, and adjusting the area ratio of each zone according to user attribute data.

[0016] Preferably, the home furnishing filling unit implements the filling design through a first algorithm, which matches the floor material, wall decoration and ceiling style that are compatible with the user attribute data based on the material association rules in the whole-house decoration dataset.

[0017] Preferably, the furniture layout generation module includes:

[0018] A baseline generation unit is used to generate at least one set of baselines on a floor plan after it has been filled with furniture, the baselines including furniture placement baselines and functional area boundary lines.

[0019] A model matching unit is used to filter size-fitting furniture and appliance models from the complete set dataset based on the baseline.

[0020] An automatic fill unit is used to fill the selected furniture and appliance models into the area defined by the baseline using a second algorithm, which includes a collision detection mechanism.

[0021] Preferably, the furniture placement baselines generated by the baseline generation unit include sofa baselines, bed baselines, and television baselines, and the position parameters of each baseline are set based on ergonomic data.

[0022] Preferably, the complete set dataset management module includes:

[0023] The data acquisition unit is used to collect data on apartment layouts, design styles, main and auxiliary materials, and home furnishing product combinations.

[0024] A data processing unit is used to standardize and dynamically update the data, wherein the annotation includes material compatibility tags and spatial function tags;

[0025] Database building blocks are used to build structured databases that support multimodal data association and retrieval.

[0026] Preferably, the virtual visualization and interaction module includes:

[0027] The web-based display unit is used to visualize design solutions using 3D rendering technology.

[0028] VR immersive unit, which is used to integrate virtual reality technology to realize immersive preview of design schemes and support scene roaming;

[0029] The real-time interaction unit is used to support adjustments to the partition boundaries, baselines, and model positions in the design scheme. After the adjustment, the intelligent house type processing module and furniture layout generation module are recalculated.

[0030] Preferably, the internal supply chain collaboration module includes:

[0031] The data interface unit is used to build a standardized data exchange interface to enable real-time data sharing between the design system and the internal supply chain management system.

[0032] The process collaboration unit is used to automatically generate a bill of materials containing furniture and main and auxiliary material information after the design scheme is confirmed, and to trigger the inventory verification process.

[0033] Preferably, the collision detection mechanism of the automatic filling unit is used to identify spatial conflicts between furniture and appliance models and walls, doors and windows, and to automatically adjust the model position or replace the adapted model based on the conflict results.

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] By integrating multimodal data through a complete home furnishing dataset, data silos are eliminated, improving data reuse and related retrieval efficiency. The intelligent apartment layout processing and furniture arrangement generation module uses algorithms to automatically divide the apartment layout, fill in furniture, and accurately arrange furniture, reducing manual intervention and improving design efficiency. Personalized design is achieved by combining user attribute data, improving the adaptability of the solution to user needs. Virtual visualization and interactive functions enhance users' perception and participation in the design solution, helping to facilitate rapid decision-making. The internal supply chain collaboration module connects the design and supply chain processes, ensuring the feasibility of the solution, optimizing overall operational efficiency, and forming a closed loop from data to design to implementation, comprehensively improving the intelligence and personalization level of home design. Attached Figure Description

[0036] Figure 1 A schematic diagram of the structure of a multimodal home space design system based on a complete home furnishing dataset provided in an embodiment of the present invention;

[0037] Figure 2 This is a schematic diagram of the structure of the complete assembly dataset management module provided in an embodiment of the present invention;

[0038] Figure 3 This is a schematic diagram of the intelligent apartment layout processing module provided in an embodiment of the present invention;

[0039] Figure 4 This is a schematic diagram of the structure of the furniture layout generation module provided in an embodiment of the present invention;

[0040] Figure 5 This is a schematic diagram of the structure of the virtual visualization and interaction module provided in an embodiment of the present invention;

[0041] Figure 6 This is a schematic diagram of the internal supply chain collaboration module provided in an embodiment of the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Please see Figures 1-6 This invention provides a multimodal home space design system based on a complete home furnishing dataset, applied to intelligent floor plan recognition, scheme generation, and virtual visualization in the home decoration industry, achieving deep collaboration with supply chain management systems. The system includes:

[0044] The integrated home furnishing dataset management module 11 is used to build and manage the integrated home furnishing dataset, which is a structured database containing apartment type data, design style data, main and auxiliary material data, and home furnishing product matching data.

[0045] In this embodiment, the integrated home furnishing dataset management module 11 refers to a functional module used to systematically build, maintain, and call the integrated home furnishing dataset. The integrated home furnishing dataset refers to a structured database containing apartment layout data, design style data, main and auxiliary material data, and home furnishing product matching data. Apartment layout data refers to information describing the spatial structure of a house (such as dimensions and wall distribution in CAD drawings and house photos). Design style data refers to parameters reflecting the characteristics of different decoration styles (such as color matching and material preferences in Nordic style). Main and auxiliary material data refers to material information used for decoration, such as 3D models of furniture, lighting parameters, and material texture images. Home furnishing product matching data refers to the combination rules between different furniture and building materials (such as the style matching relationship between sofas and coffee tables). The purpose of this module is to provide unified, high-quality data support for the entire design system. By integrating multimodal data, it eliminates data silos, achieves efficient data reuse and related retrieval, thereby improving the accuracy and generation efficiency of design schemes. Its implementation involves first collecting various types of raw data through the data acquisition unit 111, then performing standardization processing (such as labeling and classification) by the data processing unit 112, and finally forming structured storage through the database construction unit 113 to ensure that the data can be quickly called by the algorithm.

[0046] In addition, it should be noted that the data acquisition unit 111 can collect floor plans, customer-confirmed design schemes, and furniture model parameters provided by suppliers from the company's historical service cases; the data processing unit 112 can use automated annotation tools to add tags to the data (such as "small apartment" and "solid wood furniture"); and the database construction unit 113 can use distributed storage technology to support parallel access to massive amounts of data.

[0047] For example, in one specific implementation, the whole-house data set management module 11 collects the apartment type data of more than 3,000 customers served in the past three years, covering various apartment types such as two-bedroom and three-bedroom apartments. At the same time, it integrates the main and auxiliary material data of 50 design styles (such as the fabric sofa model of modern minimalist style and the parameters of metal lamps of light luxury style), and establishes the relationship between "apartment type-style-furniture" through the data processing unit 112, so that the design system can quickly match the appropriate furniture model when generating the scheme.

[0048] The intelligent apartment layout processing module 12 is used to acquire apartment layout data and divide it into zones. Based on user attribute data, it performs home furnishing design on the divided apartment layout to achieve integrated molding.

[0049] In this embodiment, the intelligent apartment layout processing module 12 refers to a functional module that automates the processing of apartment layout data, including two core steps: zoning and furnishing. Zoning refers to the process of dividing the overall apartment layout into different sub-areas according to functional requirements (such as public areas and private areas). User attribute data refers to information describing the user's family composition and usage needs (such as "a family of three including a child" or "an adult male living alone"). Furnishing design refers to the process of matching basic decoration elements such as flooring, walls, and ceilings to each sub-area after zoning. Integrated molding refers to ensuring that the decoration style and material selection of each area remain coordinated and unified. The purpose of this module is to replace the traditional manual apartment layout analysis process, achieving rapid analysis and personalized filling of the apartment layout through algorithms, improving the efficiency of the early design phase, and ensuring that the design scheme meets the actual usage needs of the user. Its implementation idea is to first perform functional zoning based on spatial characteristics (such as lighting and circulation), then adjust the zoning ratio and decoration elements in combination with user attribute data, and finally generate a complete basic decoration scheme.

[0050] Additionally, it should be noted that the cutting and zoning can use image recognition technology to extract the positions of doors, windows, and walls in the floor plan, and automatically divide the area by combining it with preset functional zoning rules (such as the living room needing to be close to the entrance door); the furniture filling can call the appropriate materials in the whole-house decoration dataset according to user attribute data (such as prioritizing environmentally friendly latex paint and non-slip flooring for children's rooms).

[0051] For example, in one specific implementation, after receiving photos of an 80-square-meter two-bedroom apartment uploaded by a user, the intelligent apartment layout processing module 12 determines the room's lighting distribution through image recognition, divides the south-facing room into a children's room (accounting for 20% of the area), divides the north-facing room into a master bedroom, and connects the living room and dining room into a public area; at the same time, based on the user attribute data of "a family of three including a 6-year-old child", it fills the children's room with non-slip flooring and anti-collision wall materials, and fills the living room with modern-style floor tiles and ceilings, so as to achieve a unified style and functional compatibility in each area.

[0052] The furniture layout generation module 13 is used to generate baselines on the floor plan after furniture filling, and to fill furniture and appliance models based on the baselines.

[0053] In this embodiment, the furniture layout generation module 13 refers to the functional module for arranging furniture and appliances on the floor plan after basic decoration is completed. Here, the baseline refers to the reference line used to standardize the placement of furniture (such as the sofa baseline, bed baseline, etc.), and the furniture and appliance models refer to the three-dimensional digital models retrieved from the complete decoration dataset (such as the three-dimensional models of sofas, televisions, and refrigerators). The purpose of this module is to achieve precise furniture layout through algorithms, ensuring that the furniture placement conforms to the principles of ergonomics and maximizing space utilization, while maintaining consistency with the basic decoration style. The implementation approach is to first generate baselines based on space function and ergonomics, then select models with suitable sizes and styles from the dataset based on the baselines, and finally ensure that the layout is conflict-free through a collision detection algorithm.

[0054] Additionally, it should be noted that baseline generation can be combined with room dimensions and usage scenarios (e.g., the baseline length of the sofa in the living room is 70% of the wall width); model selection can be achieved through style tag matching (e.g., a Nordic-style sofa paired with a coffee table of the same style); and collision detection can automatically identify and adjust the position of furniture that overlaps with walls, doors, and windows.

[0055] For example, in one specific implementation, the furniture layout generation module 13 generates a sofa baseline 40cm from the back wall and 2.8m long in the living room area after basic decoration is completed, and a TV baseline parallel to the sofa baseline and 3.5m apart; then it retrieves a 2.7m long Nordic-style three-seater sofa and a 55-inch TV model from the complete set dataset, places them along the baseline, and confirms through collision detection that the sofa and the entrance door are not obstructed, thus completing the living room furniture layout.

[0056] Virtual visualization and interaction module 14 is used to display design schemes in the form of 3D models and supports interactive operations on the design schemes.

[0057] In this embodiment, the virtual visualization and interaction module 14 refers to a functional module that displays design schemes in a visual form and supports user adjustments. The 3D model refers to a three-dimensional digital model constructed using computer technology (including details such as floor plan, furniture, and materials), and the interactive operation refers to user modifications to the design scheme (such as adjusting furniture positions or changing materials). The purpose of this module is to provide users with an intuitive and immersive preview experience of the scheme, while supporting real-time adjustments to meet personalized needs, thereby improving customer satisfaction and decision-making efficiency. Its implementation involves converting abstract design data into a perceptible 3D scene using web and VR technologies, then receiving user operations through an interactive interface and triggering system recalculation to update the scene display in real time.

[0058] Additionally, it should be noted that the web-based display can utilize lightweight rendering technology to ensure fast loading on ordinary devices; the VR immersive preview supports head-mounted displays, providing 360° panoramic views; interactive operations can be achieved through mouse drag-and-drop, voice commands, etc., with a system response time of ≤2 seconds.

[0059] For example, in one specific implementation, the virtual visualization and interaction module 14 converts the generated design scheme into a three-dimensional model, which can be loaded and viewed by web users within 15 seconds; users wearing VR devices can immerse themselves in the room and experience the actual spatial scale; when a user moves the sofa position 50cm to the right by dragging, the system automatically recalculates the surrounding traffic flow and updates the three-dimensional model to ensure real-time feedback of the adjustment results.

[0060] Internal supply chain collaboration module 15 is used to realize data sharing and process collaboration between the design system and the internal supply chain management system.

[0061] In this embodiment, the internal supply chain collaboration module 15 refers to a functional module that enables data exchange and process linkage between the design system and the internal supply chain management system. The internal supply chain management system is a system used to manage enterprise procurement, inventory, production, and other aspects. Data sharing refers to the exchange of information between the design system and the supply chain system (such as furniture inventory status and production cycle). Process collaboration refers to the automatic triggering of supply chain operations (such as generating a purchase list and arranging production) after the design scheme is confirmed. The purpose of this module is to break down the barriers between design and the supply chain, ensuring that the furniture and building materials in the design scheme can be produced, shortening the design-to-construction cycle, and improving overall operational efficiency. Its implementation involves achieving real-time information synchronization through standardized data interfaces and then converting the design results into supply chain instructions through preset process rules.

[0062] Additionally, it should be noted that the data interface can use a RESTful API, supporting real-time comparison of material information in the design scheme with supply chain inventory data; after the design scheme is confirmed, the process collaboration can automatically generate a bill of materials containing furniture models and quantities, and push it to the procurement system to trigger the replenishment process.

[0063] For example, in one specific implementation, the internal supply chain collaboration module 15 obtains the information of 10 furniture models included in the design scheme through a data interface. After comparing it with the inventory data of the internal supply chain management system, it finds that 2 pieces of furniture are in short supply. Subsequently, a bill of materials is automatically generated, in which the 8 pieces of furniture with sufficient inventory trigger the outbound process, and the 2 pieces of furniture with insufficient inventory trigger the purchase order, thereby achieving seamless connection between the design scheme and the supply chain.

[0064] In an optional embodiment, the complete set dataset management module 11 includes:

[0065] Data acquisition unit 111 is used to collect apartment layout data, design style data, main and auxiliary material data, and home furnishing product matching data;

[0066] Data processing unit 112 is used to standardize and dynamically update the data, wherein the annotation includes material compatibility tags and spatial function tags;

[0067] Database building unit 113 is used to build a structured database that supports multimodal data association retrieval.

[0068] In this embodiment, the data acquisition unit 111 refers to the unit that collects various design-related data. This includes: floor plan data (descriptions of house structure such as CAD drawings and scanned photos); design style data (parameters reflecting style characteristics such as the proportion of metal elements and color saturation in a "light luxury" style); main and auxiliary material data (details of furniture and building materials such as sofa dimensions, materials, and prices; flooring wear resistance and environmental protection levels); and home furnishing product matching data (rules for combining different products such as "style consistency between coffee table and sofa" and "size matching of dining table and chairs"). The purpose of this unit is to provide the system with rich and comprehensive raw data. The data processing unit 112 refers to the unit that processes the collected data. This includes: standardized labeling (adding uniformly formatted tags such as "material: solid wood" and "style: Chinese"); dynamic updating (regularly adding new data such as new customer cases and new furniture models); material compatibility tags (labels indicating applicable scenarios for materials such as "anti-slip - bathroom" and "wear-resistant - living room"); and space function tags (labels indicating applicable spaces for products such as "bedroom - bed" and "living room - sofa"). The purpose of this unit is to improve data quality and facilitate efficient retrieval and access. Database building unit 113 refers to the unit for constructing data storage structures. Supporting multimodal data association retrieval means that other types of data can be queried through association using one data type (e.g., simultaneously retrieving corresponding apartment layout, furniture, and material data using the "Nordic style" tag). The purpose of this unit is to achieve structured data storage and efficient reuse.

[0069] Additionally, it should be noted that the data acquisition unit 111 can automatically synchronize the furniture model data of the supplier through the API interface; the data processing unit 112 can use natural language processing technology to automatically extract material requirement tags (such as "easy to clean") from user reviews; and the database construction unit 113 can use a graph database to store data relationships (such as using "apartment type" as a node and associating it with sub-nodes such as "style" and "furniture").

[0070] For example, in one feasible implementation, the data acquisition unit 111 collects CAD drawings of 5,000 apartment types, feature parameters of 50 styles, 3D models of 3,000 pieces of furniture, and matching examples; the data processing unit 112 adds tags such as "material compatibility: living room / bedroom" and "space function: three-seater sofa" to each piece of furniture, and updates 100 new case data sets every month; the graph database constructed by the database construction unit 113 supports the return of suitable apartment structures, furniture models, and material combinations through the search criteria of "modern style + two-bedroom", with a response time of ≤1 second.

[0071] In an optional embodiment, the smart apartment layout processing module 12 includes:

[0072] Apartment type zoning unit 121 is used to cut and zon the apartment type data using a spatial semantic segmentation algorithm. The zoning is based on lighting conditions, circulation distance and functional attributes to determine the zoning boundaries.

[0073] User attribute parsing unit 122 is used to receive user attribute data, which includes family member composition and usage needs;

[0074] The home furnishing filling unit 123 is used to fill the floor, walls and ceiling of the divided floor plan according to the user attribute data and the whole-house furnishing dataset, so that the space layout is full and the functions are complete.

[0075] In this embodiment, the apartment layout zoning unit 121 refers to a unit that uses a spatial semantic segmentation algorithm to functionally partition apartment data. The spatial semantic segmentation algorithm refers to deep learning-based image segmentation technology (such as the U-Net model), which can identify features such as walls, doors, and windows in the apartment layout and divide them into areas. Lighting conditions refer to the intensity and duration of natural light received by the rooms. Circulation distance refers to the walking path length from the entrance door to each area. Functional attributes refer to the intended use of each area (such as rest, activity, cooking, etc.). This unit determines the zoning boundaries using these parameters to achieve rational space planning and improve space utilization. The user attribute parsing unit 122 refers to a unit that receives and processes user attribute data. Family member composition refers to information such as the age, gender, and number of residents (e.g., "2 adults and 1 child," "including a 6-year-old child"). Usage needs refer to the user's specific requirements for the space (e.g., "needs a study," "children's room needs storage space"). Its purpose is to transform user needs into design parameters, providing a basis for personalized design. Home Furnishing Filling Unit 123 refers to the unit that performs basic decoration design based on user attribute data and the complete decoration dataset. The filling design for floors, walls, and ceilings refers to matching materials (such as flooring, wallpaper, and ceiling styles) to each area to ensure a full and functional space. This means ensuring a unified decoration style and meeting usage needs in each area. The implementation idea of ​​this unit is to combine user attributes with material association rules in the dataset to automatically match suitable decoration elements, improving the personalization and efficiency of the design.

[0076] Additionally, it should be noted that in one possible implementation, the apartment zoning unit 121 can divide the area into zones by a preset threshold (e.g., areas with a daylight factor ≥ 300 lux are divided into public activity areas); the user attribute analysis unit 122 can convert the requirement of "including the elderly" into the parameter of "the bathroom needs to be designed with anti-slip properties"; and the furniture filling unit 123 can prioritize matching environmentally friendly materials (e.g., water-based paint, E0 grade boards) for children's rooms.

[0077] For example, in one feasible implementation, the apartment zoning unit 121 uses the U-Net model to divide the 80㎡ apartment floor plan, dividing the south-facing area with a daylight factor ≥350lux into the living room (public activity area), and dividing the area ≤5m from the entrance door into the kitchen (core functional area); the user attribute parsing unit 122 receives the attribute data of "1 adult and 1 child (5 years old)" and parses out the need for "children's activity space"; the furniture filling unit 123 fills the living room with wear-resistant floor tiles and the children's room with non-slip flooring and cartoon wallpaper according to the need and the whole-house decoration dataset, so as to match the space function with the user's needs.

[0078] In an optional embodiment, the zoning rules of the apartment zoning unit 121 include: dividing areas with a daylight factor ≥ a preset value into public activity areas, dividing areas with a circulation distance ≤ a preset value into core functional areas, and adjusting the area ratio of each zone according to user attribute data.

[0079] In this embodiment, the zoning rules refer to the specific standards for dividing the areas in the apartment zoning unit 121. The daylight factor refers to an indicator measuring light intensity (in lux), the preset value refers to the system-set daylight factor threshold (e.g., 300 lux), the public activity area refers to areas where family members gather and engage in activities (e.g., living room, dining room), the circulation distance refers to the walking path length from the entrance door to the area, and the core functional area refers to frequently used areas that require convenient access (e.g., kitchen, bathroom). The area ratio of each zone refers to the ratio of the area of ​​each zone to the total area of ​​the apartment. The purpose of these rules is to ensure that the zoning conforms to ergonomics and living habits, improving living comfort. The implementation approach involves determining the basic zoning through quantitative parameters (daylight, circulation), and then dynamically adjusting the area based on user attribute data (e.g., "multi-person families" need to expand the public area), making the zoning more closely match actual needs.

[0080] It should be noted that the preset values ​​can be adjusted according to the regional lighting conditions (e.g., the preset daylight factor is 250 lux in northern regions and 350 lux in southern regions); the preset distance for the circulation path in the core functional area can be set to 5m to ensure the convenience of activities such as cooking and washing; when the user attribute is "elderly living alone", the living room area can be reduced and the bedroom area can be increased to 30% of the total area. For example, in a feasible implementation, the unit 121 divides the area with a daylight factor ≥300 lux into the living room (accounting for 25%) and the dining room (accounting for 15%), and the area ≤5m from the entrance door into the kitchen (accounting for 10%) and the bathroom (accounting for 8%); for the user attribute of "family of three", the proportion of the children's room area is adjusted from 12% to 18% to ensure sufficient activity space.

[0081] In another possible implementation, when the smart apartment layout processing module 12 acquires structured data with pre-divided areas (the pre-divided areas refer to the apartment layout data where the boundaries of each functional area are clearly marked, such as the division of the living room, bedroom, and kitchen defined by coordinates or labels), the module can skip the zoning step of the apartment layout zoning unit 121 and directly enter the furniture filling stage. At this time, the apartment layout zoning unit 121 is in a dormant state, while the user attribute parsing unit 122 still receives and processes user attribute data (such as family member composition and usage preferences) and directly transmits the parsing results to the furniture filling unit 123. The furniture filling unit 123, based on the area-material association rules in the integrated decoration dataset (such as "bedrooms are matched with silent flooring by default, and living rooms are matched with wear-resistant floor tiles"), combined with user attribute data (such as "elderly residents" corresponding to the bedroom having added anti-slip attributes), performs floor, wall, and ceiling filling design for each divided area, while ensuring the consistency and integration of the overall design through a cross-area style verification mechanism (such as unifying the color saturation deviation of each area to ≤10%).

[0082] For example, in one feasible implementation, the apartment layout processing module 12 receives apartment layout data that has been pre-divided into three areas: living room, master bedroom, and children's room (boundaries clearly defined by coordinates), and the user attribute data is "elderly couple + occasional grandchildren living there". After the user attribute parsing unit 122 parses out the requirement that "the bedroom needs to be non-slip and the living room needs to be spacious", the furniture filling unit 123 directly calls the adaptation rules in the integrated decoration dataset: fill the living room with 600×600mm light gray floor tiles (to enhance the sense of spaciousness), fill the master bedroom with non-slip solid wood flooring and match it with anti-collision wall material, and fill the children's room (temporary use) with replaceable cartoon wallpaper and play mats. The style verification ensures that the warm color deviation of the three is ≤5%. Finally, the integrated design that meets the requirements is completed without re-dividing the areas, and the design cycle is shortened by 30% compared with the data processing without partitioning.

[0083] In an optional embodiment, the home furnishing filling unit 123 implements the filling design through a first algorithm, which matches the floor material, wall decoration and ceiling style that are compatible with the user attribute data based on the material association rules in the whole-house furnishing dataset.

[0084] In this embodiment, the first algorithm refers to the algorithm used by the home furnishing filling unit 123 to match decoration materials. The material association rules in the whole-house decoration dataset refer to the matching logic between different materials (e.g., "Nordic style living rooms are often paired with light-colored wood flooring + white ceilings"), while user attribute data adaptation refers to the matching of material selection with user needs (e.g., "families with pets are suited to wear-resistant flooring"). The purpose of this algorithm is to achieve fast and accurate material matching, ensuring a unified decoration style and meeting the actual needs of users. Its implementation involves first extracting material association rules from the whole-house decoration dataset (e.g., mining the correspondence between "style-space-material" through association analysis), then combining this with user attribute data (e.g., family members, usage habits) to filter suitable material combinations, ultimately completing the filling design.

[0085] For example, in one feasible implementation, the first algorithm analyzes 1,000 Nordic style cases in the whole-house decoration dataset and mines the association rule of "living room - light-colored wood flooring - white plaster ceiling". Combined with the user's attribute of "having a 6-year-old child", it matches the children's room area with laminate flooring with a wear resistance coefficient ≥ AC4 and anti-collision wallpaper, which not only meets the requirements of style unity, but also meets the safety requirements of children.

[0086] Additionally, it should be noted that the first algorithm can be a decision tree-based matching algorithm, using "style," "space type," and "user attributes" as feature variables to output corresponding material schemes; or it can be a deep learning algorithm (such as CNN), which learns material matching rules through a training dataset to achieve end-to-end material recommendations. For example, when the user attribute is "member with allergies," the algorithm will prioritize recommending formaldehyde-free wallpaper and mite-proof flooring.

[0087] For example, in one feasible implementation, the first algorithm is a rule-based reasoning decision tree algorithm, which achieves material matching through the following steps:

[0088] Feature extraction: Extract key features from user attribute data, including family composition (e.g., "including 6-year-old child" or "adult living alone"), usage needs (e.g., "storage priority" or "easy to clean"), and style preferences (e.g., "Nordic" or "Chinese").

[0089] Rule base construction: Based on the material association rules in the complete set dataset, a three-level decision tree is constructed:

[0090] First-level node: Filter material categories by style preference (e.g., Nordic style corresponds to "light-colored wood and fabric");

[0091] Secondary nodes: refine material properties according to spatial function (such as living room, bedroom, children's room) (e.g., children's room needs to be "environmentally friendly and non-slip");

[0092] Level 3 nodes: Specific parameters are determined according to the user's special needs (e.g., "easy to clean" corresponds to "glazed tiles, waterproof wallpaper").

[0093] Matching inference: Input user features into the decision tree, and match rules step by step:

[0094] If the user's attributes are "Nordic style + family of three including children + living room", the first-level match is "light-colored wood", the second-level match is "wear-resistant material for living room", and the third-level match is "scratch-resistant flooring".

[0095] Output flooring materials (such as "ash wood laminate flooring"), wall decorations (such as "off-white latex paint"), and ceiling styles (such as "plaster molding ceiling").

[0096] Consistency check: Call the material compatibility rules in the complete set dataset (such as "the color contrast between the floor and the wall is ≤30%) to check the matching results. If the match is not met, backtrack to the second-level node to select again.

[0097] In an optional embodiment, the furniture layout generation module 13 includes:

[0098] The baseline generation unit 131 is used to generate at least one set of baselines on the floor plan after it has been filled with furniture, the baselines including furniture placement baselines and functional area boundary lines.

[0099] Model matching unit 132 is used to filter size-fitting furniture and appliance models from the complete set dataset based on the baseline.

[0100] An automatic filling unit 133 is used to fill the selected furniture and appliance models into the area defined by the baseline using a second algorithm, the second algorithm including a collision detection mechanism.

[0101] In this embodiment, the baseline generation unit 131 refers to the unit that generates reference lines on the floor plan. Baselines are lines that standardize furniture placement (e.g., the back edge of a sofa against a wall, the center line of a bed), furniture placement baselines are positional reference lines set for specific furniture, and functional area boundary lines are lines that divide different functional sub-areas (e.g., the dividing line between the living room and dining room). The purpose of this unit is to provide precise positioning for furniture layout, ensuring reasonable placement. The model matching unit 132 refers to the unit that filters furniture and appliance models based on the baselines. Size adaptation refers to the matching of furniture dimensions with the space defined by the baselines (e.g., sofa length ≤ sofa baseline length). The purpose of this unit is to quickly filter models that match the space dimensions and style from the complete set dataset, improving layout efficiency. The automatic filling unit 133 refers to the unit that places the filtered models into designated areas. The second algorithm refers to the algorithm for automatic layout, and the collision detection mechanism refers to the technology for identifying overlaps between models and walls, doors, windows, or other model spaces. The purpose of this unit is to ensure that furniture placement is conflict-free, improving the rationality of the layout.

[0102] Additionally, it should be noted that the baseline generation unit 131 can dynamically adjust the baseline length according to the space dimensions (e.g., in a 3m wide living room, the sofa baseline length is set to 2.2m); the model matching unit 132 can combine style tags for filtering (e.g., matching a metal-framed sofa to a "modern style living room"); the second algorithm can be a layout algorithm based on a greedy strategy, placing large furniture (e.g., sofa, bed) first, and then arranging smaller furniture (e.g., coffee table, bedside table). For example, in a feasible implementation, the baseline generation unit 131 generates a sofa baseline 40cm from the back wall and 2.5m long in the living room, and a TV baseline parallel to the sofa and 3m apart; the model matching unit 132 filters out a 2.4m long modern style sofa and a 55-inch TV model from the complete set dataset; the automatic filling unit 133 places the sofa along the baseline using the second algorithm, hangs the TV at the TV baseline position, and at the same time, the collision detection mechanism detects a conflict between the coffee table and the aisle, automatically shifting it 30cm towards the sofa to complete a conflict-free layout.

[0103] In another feasible implementation, the second algorithm is an improved A* algorithm based on a greedy strategy, which achieves furniture filling and collision detection through the following steps:

[0104] Baseline analysis: Obtain the coordinate parameters of the furniture placement baseline (such as the starting point (x1, y1) and ending point (x2, y2) of the sofa baseline) from the baseline generation unit 131, and calculate the baseline length and the boundary of the area where it is located;

[0105] Model sorting: From the furniture models filtered by model matching unit 132, sort them by size (e.g., sofa > coffee table > side table), and prioritize the processing of large furniture;

[0106] Initial layout:

[0107] For the sofa model, align it with the midpoint of the baseline and calculate its three-dimensional bounding box coordinates (xmin, ymin, zmin) - (xmax, ymax, zmax).

[0108] Call the collision detection mechanism to determine the distance between the model and walls, doors and windows through spatial intersection calculation (must be ≥10cm). If the distance is satisfied, fix the position.

[0109] Iterative filling:

[0110] Based on the positions of the already placed furniture, the A* algorithm is used to plan the path of the remaining furniture (e.g., the coffee table needs to be located directly in front of the sofa), and candidate placement coordinates are generated;

[0111] For each candidate coordinate, calculate the minimum distance to the already placed furniture (e.g., the distance between the coffee table and the sofa is ≥50cm), and eliminate overlapping areas through collision detection;

[0112] Conflict Correction:

[0113] If a conflict is detected (such as a coffee table overlapping with a passageway), fine-tune the candidate coordinates along the baseline (moving 5cm each time) until the conflict is eliminated;

[0114] If the conflict persists after fine-tuning, then use the complete dataset to filter out smaller models of the same style (e.g., replace the 120cm coffee table with a 90cm one), and repeat steps 3-4.

[0115] Layout optimization: After all furniture is placed, calculate the overall space utilization rate (≥60%). If it does not meet the standard, readjust the position of large furniture until the requirements are met.

[0116] In an optional embodiment, the furniture placement reference lines generated by the reference line generation unit 131 include sofa reference lines, bed reference lines, and television reference lines, and the position parameters of each reference line are set based on ergonomic data.

[0117] In this embodiment, the sofa baseline refers to a reference line used to determine the sofa's placement, the bed baseline refers to a reference line used to determine the bed's placement, and the television baseline refers to a reference line used to determine the television's installation position. Ergonomic data refers to parameters based on human physiological characteristics and range of motion (e.g., "the optimal viewing distance between the sofa and the television is 2.5-3 times the diagonal of the television"). Positional parameters refer to parameters such as the distance and length of the baselines (e.g., the sofa baseline is 30-50cm from the back wall). The purpose of this unit is to ensure that furniture placement conforms to ergonomics and improves user comfort. The implementation involves converting ergonomic data into specific positional parameters (e.g., calculating the distance between the sofa and the television based on the television size), and then combining this with spatial dimensions to generate baselines, providing a scientific basis for furniture layout.

[0118] Additionally, it should be noted that the sofa baseline length can be set to 70%-80% of the wall width (e.g., a 4m wide wall corresponds to a 2.8-3.2m baseline), the bed baseline can be arranged along the longest side of the room (ensuring the headboard is against the wall), and the TV baseline height can be set to 1.2-1.5m (suitable for seated viewing distance). For special apartment layouts (such as long and narrow living rooms), the baseline position parameters can be adjusted appropriately (e.g., shortening the distance between the sofa and the TV to twice the diagonal length).

[0119] For example, in one feasible implementation, the baseline generation unit 131 generates a sofa baseline with a length of 2.8m (80% of the wall width) and a distance of 40cm from the back wall for a living room that is 3.5m wide; for a bedroom that is 1.8m wide, it generates a bed baseline along the long side of 3.2m (10cm from the headboard to the wall); and based on a 55-inch TV (1.4m diagonal), it generates a TV baseline with a distance of 3.5m (2.5 times the diagonal) from the sofa baseline. All parameters conform to ergonomic data to ensure comfortable use.

[0120] In an optional embodiment, the collision detection mechanism of the automatic filling unit 133 is used to identify spatial conflicts between furniture and appliance models and walls, doors and windows, and to automatically adjust the model position or replace the adapted model based on the conflict results.

[0121] In this embodiment, the collision detection mechanism refers to the technology used by the automatic filling unit 133 to identify spatial conflicts. Spatial conflicts between furniture and appliance models and walls, doors, and windows refer to the overlap between the three-dimensional boundaries of the model and the boundaries of the walls, doors, and windows (e.g., a sofa partially embedded in a wall). The purpose of this mechanism is to ensure that the furniture placement is reasonable and to avoid situations where it cannot be actually installed. The implementation involves calculating the intersection of the three-dimensional coordinates of the model and spatial elements (walls, doors, and windows). If the intersection exceeds a preset threshold (e.g., 5cm), it is determined to be a conflict. Automatic adjustment based on the conflict result means that the system eliminates the conflict by moving the model's position (e.g., moving the sofa away from the wall) or replacing it with a smaller model (e.g., replacing a three-seater sofa with a two-seater sofa), ensuring the feasibility of the layout.

[0122] Additionally, it should be noted that in one possible implementation, the collision detection mechanism can use a bounding box algorithm (such as an AABB axis-aligned bounding box) to quickly calculate spatial intersection. The preset threshold can be adjusted according to the model size (e.g., a threshold of 10cm for large furniture and 5cm for small furniture). The adjustment strategy can prioritize moving the model (e.g., translating along a baseline), and only replace the model if moving fails (selecting a smaller model of the same style from the complete set dataset). For example, in a feasible implementation, when the automatic filling unit 133 is arranging a 2.8m long sofa, the collision detection mechanism uses the bounding box algorithm to find that the right side of the sofa overlaps with the door frame by 15cm (exceeding the 10cm threshold), which is determined to be a conflict. The system first tries to translate the sofa 20cm to the left, but finds that it still overlaps with the left wall by 5cm. Finally, a 2.5m long sofa model of the same style is selected from the complete set dataset, and after replacement, there is no conflict, and the layout is completed.

[0123] In an optional embodiment, the virtual visualization and interaction module 14 includes:

[0124] Web-based display unit 141 is used to visualize the design scheme using 3D rendering technology;

[0125] VR immersive unit 142 is used to integrate virtual reality technology to realize immersive preview of design schemes and support scene roaming;

[0126] The real-time interaction unit 143 is used to support the adjustment of partition boundaries, baselines and model positions in the design scheme. After the adjustment, the intelligent house type processing module 12 and the furniture layout generation module 13 are triggered to recalculate.

[0127] In this embodiment, the Web-based display unit 141 refers to the unit that displays the design scheme through a web page. The 3D rendering technology refers to the technology that converts 3D models into visual images (such as WebGL), aiming to allow users to quickly view the design effect on ordinary devices. The VR immersive unit 142 refers to the unit that displays the scheme through virtual reality technology. Immersive preview refers to users gaining an immersive viewing experience through VR devices, and scene roaming refers to users being able to freely move their viewpoint within a virtual scene (such as moving from the living room to the bedroom). The purpose of this unit is to improve the user experience and make the design effect more intuitive. The real-time interaction unit 143 refers to the unit that supports users in modifying the design scheme. Adjusting partition boundaries, baselines, and model positions means that users can drag boundary lines to change the size of the area, move baselines to adjust the position of furniture, and drag models to change the placement angle, etc. Recalculation triggered after adjustment means that the system automatically updates related design elements (such as automatically adapting the coffee table position after moving the sofa baseline). The purpose of this unit is to meet users' personalized needs and improve the flexibility of the scheme.

[0128] Additionally, it should be noted that in one possible implementation, the Web-based display unit 141 may employ LOD (Level of Detail) technology to automatically adjust rendering precision based on device performance; the VR immersive unit 142 supports mainstream VR devices (such as Oculus Quest) with a frame rate ≥60fps to avoid motion sickness; and the real-time interaction unit 143 can be adjusted via voice commands (such as "move the sofa 50cm to the right") with a system response time ≤2 seconds.

[0129] For example, in one feasible implementation, the Web-based display unit 141 uses WebGL technology to render the design scheme, which can be loaded and viewed 360° by users on ordinary office computers within 10 seconds; the VR immersive unit 142 allows users to wear VR glasses to roam the virtual space and experience the actual scale of the room layout; when the user expands the living room partition boundary by 1m by dragging, the real-time interaction unit 143 triggers the floor plan partitioning unit 121 to recalculate, the sofa baseline is automatically extended, and the model matching unit 132 re-selects a longer sofa to match, realizing the linkage update of the scheme.

[0130] In an optional embodiment, the internal supply chain collaboration module 15 includes:

[0131] Data interface unit 151 is used to build a standardized data exchange interface to realize real-time data sharing between the design system and the internal supply chain management system;

[0132] The process collaboration unit 152 is used to automatically generate a bill of materials containing furniture and main and auxiliary material information after the design scheme is confirmed, and to trigger the inventory verification process.

[0133] In this embodiment, the data interface unit 151 refers to the unit that realizes data exchange between systems. The standardized data exchange interface refers to an interface following a unified format (such as RESTful API). The design system refers to the multimodal home space design system of this application. The internal supply chain management system refers to the enterprise's system for managing procurement, inventory, and production (such as an ERP system). Real-time data sharing refers to the real-time synchronization of material information (such as furniture models and quantities) in the design scheme with inventory and price data in the supply chain system. The purpose of this unit is to break down data barriers and ensure information consistency. The process collaboration unit 152 refers to the unit that realizes the linkage between design and supply chain processes. Design scheme confirmation refers to the final design scheme determined by the user. The bill of materials refers to a list containing the models, specifications, and quantities of furniture and main and auxiliary materials. The inventory verification process refers to checking whether the inventory of materials in the supply chain system meets the scheme requirements (if the inventory is insufficient, procurement is triggered). The purpose of this unit is to connect the entire process from design to production and shorten the delivery cycle.

[0134] Additionally, it should be noted that in one possible implementation, the data interface unit 151 can use encrypted transmission to ensure data security and support high-frequency data synchronization of 100 times per second; the process collaboration unit 152 can add priority identifiers (such as "furniture urgently needed for installation") to the bill of materials, and the inventory verification process can automatically distinguish between spot and pre-sale materials, triggering delivery and production instructions respectively.

[0135] For example, in one feasible implementation, the data interface unit 151 synchronizes information such as "three-seater sofa (model A123)" and "solid wood flooring (model B456)" in the design scheme to the internal supply chain management system in real time through a RESTful API to obtain the current inventory data (A123 is in stock, B456 is out of stock); after the design scheme is confirmed, the process collaboration unit 152 automatically generates a bill of materials, triggers the inventory verification process, arranges the shipment of A123, and triggers a purchase order for B456, so as to achieve seamless connection between design and supply chain.

[0136] In this embodiment, multimodal data is integrated through a complete set dataset to eliminate data silos and improve data reuse and related retrieval efficiency; the intelligent apartment type processing and furniture layout generation module 13 uses algorithms to automatically divide the apartment type into zones, fill in furniture, and accurately arrange furniture, reducing manual intervention and improving design efficiency; personalized design is achieved by combining user attribute data to improve the adaptability of the solution to user needs; virtual visualization and interactive functions enhance users' perception and participation in the design solution, helping to make quick decisions; the internal supply chain collaboration module 15 connects the design and supply chain processes to ensure the feasibility of the solution, optimize overall operational efficiency, form a closed loop from data to design to implementation, and comprehensively improve the intelligence and personalization level of home design.

[0137] It should be noted that the above examples are merely specific embodiments of the present invention, and the present invention is obviously not limited to the above embodiments, with many similar variations. All modifications that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should fall within the protection scope of this invention.

[0138] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multimodal home space design system based on a complete home furnishing dataset, characterized in that, include: The integrated home furnishing dataset management module is used to build and manage integrated home furnishing datasets, which are structured databases containing apartment layout data, design style data, main and auxiliary material data, and home furnishing product matching data. The intelligent apartment layout processing module is used to acquire apartment layout data and divide it into zones. Based on user attribute data, it performs home furnishing design on the divided apartment layout to achieve integrated molding. The furniture layout generation module is used to generate baselines on the floor plan after furniture filling, and to fill in furniture and appliance models based on the baselines; The virtual visualization and interaction module is used to display design schemes in the form of 3D models and supports interactive operations on the design schemes. The internal supply chain collaboration module is used to enable data sharing and process collaboration between the design system and the internal supply chain management system; The intelligent apartment layout processing module includes: The apartment layout zoning unit is used to divide the apartment data into zones using a spatial semantic segmentation algorithm. The zoning is based on lighting conditions, circulation distance, and functional attributes to determine the zoning boundaries. The user attribute parsing unit is used to receive user attribute data, which includes family member composition and usage needs. The home furnishing filling unit is used to fill in the floor, walls, and ceiling of the divided floor plan based on the user attribute data and the complete home furnishing dataset, so as to make the space layout full and the functions complete; The furniture layout generation module includes: A baseline generation unit is used to generate at least one set of baselines on a floor plan after it has been filled with furniture, the baselines including furniture placement baselines and functional area boundary lines. A model matching unit is used to filter size-fitting furniture and appliance models from the complete set dataset based on the baseline. An automatic fill unit is used to fill the selected furniture and appliance models into the area defined by the baseline using a second algorithm, which includes a collision detection mechanism. The internal supply chain collaboration module includes: The data interface unit is used to build a standardized data exchange interface to enable real-time data sharing between the design system and the internal supply chain management system. The process collaboration unit is used to automatically generate a bill of materials containing furniture and main and auxiliary material information after the design scheme is confirmed, and to trigger the inventory verification process. The home furnishing filling unit implements the filling design through a first algorithm, which is based on the material association rules in the whole-house decoration dataset and matches the floor material, wall decoration and ceiling style that are compatible with the user attribute data. The virtual visualization and interaction module includes: The web-based display unit is used to visualize design solutions using 3D rendering technology. VR immersive unit, which is used to integrate virtual reality technology to realize immersive preview of design schemes and support scene roaming; The real-time interaction unit is used to support the adjustment of partition boundaries, baselines and model positions in the design scheme. After the adjustment, the intelligent house type processing module and furniture layout generation module are recalculated. The collision detection mechanism of the automatic filling unit is used to identify spatial conflicts between furniture and appliance models and walls, doors and windows, and automatically adjust the model position or replace the adapted model based on the conflict results.

2. The multimodal home space design system based on a complete home furnishing dataset according to claim 1, characterized in that, The zoning rules for the apartment layout include: dividing areas with a daylight factor greater than or equal to a preset value into public activity areas, dividing areas with a circulation distance less than or equal to a preset value into core functional areas, and adjusting the area ratio of each zone according to user attribute data.

3. The multimodal home space design system based on a complete home furnishing dataset according to claim 1, characterized in that, The baseline generation unit generates furniture placement baselines including sofa baselines, bed baselines, and television baselines, with the position parameters of each baseline set based on ergonomic data.

4. The multimodal home space design system based on a complete home furnishing dataset according to claim 1, characterized in that, The complete set dataset management module includes: The data acquisition unit is used to collect data on apartment layouts, design styles, main and auxiliary materials, and home furnishing product combinations. A data processing unit is used to standardize and dynamically update the data, wherein the annotation includes material compatibility tags and spatial function tags; Database building blocks are used to build structured databases that support multimodal data association and retrieval.

Images