A method for constructing a three-dimensional model of a gift box packaging design

Abstract

The present application relates to the technical field of gift packaging three-dimensional modeling, and discloses a kind of construction methods of gift box packaging design three-dimensional model.The method includes collecting natural language description, hand-drawing sketch image and physical material attribute parameter;Analysis natural language construction design intent graph, analyze sketch to generate geometry feature code;The knowledge alignment fusion of the two forms enhanced design semantic representation;Initial three-dimensional design framework is generated using generative design algorithm;Integrate material parameters and carry out compatibility simulation test, start parameter self-adaptive adjustment cycle according to feasibility evaluation report, dynamically optimize component parameters;Render output final model.Through the deep fusion of multi-modal design information and the preposition of physical attributes dynamic optimization, the accuracy of design intent expression and the producibility of three-dimensional model are improved, and the automation and intelligentization of design process are realized.

Description

Technical Field

[0001] This invention relates to the field of 3D modeling technology for gift packaging, specifically a method for constructing a 3D model of gift box packaging design. Background Technology

[0002] In existing gift box packaging design processes, designers' creative input methods are relatively limited. When designers attempt to combine natural language descriptions with hand-drawn sketches to express concepts, current technologies typically employ separate data processing channels. The natural language processing module is responsible for parsing text keywords, while the computer vision module focuses on recognizing the geometric contours of the sketches. This parallel processing approach leads to a disconnect between semantic and geometric information. The stylistic intent conveyed by the text is difficult to effectively correlate and verify with the specific morphological features of the sketches, resulting in a deviation between the generated 3D conceptual model and the designer's overall intent.

[0003] After the initial construction of the 3D model is completed, the integration of material properties usually occurs in the later stages of the design process. The traditional approach is to first build the model, then assign it physical material parameters, and finally test the structural strength or printing quality through limited simulations. This method is a passive verification process; if a mismatch is found between the material and the designed structure, the designer must manually return to the initial design steps for significant modifications. This iterative process is not only inefficient but also heavily reliant on the designer's personal experience, making it difficult to guarantee the feasibility and optimization of the design solution.

[0004] Current technology lacks a closed-loop system that connects multimodal creative input to physical realization verification. The deep semantics of design intent and the geometric expression of morphological features fail to achieve unified representation, and the constraints of material physical properties on the design structure are not actively and dynamically fed back and optimized in the early and later stages of design. This leads to problems such as long design iteration cycles, high manufacturability risks of design results, and inaccurate creative expression. Summary of the Invention

[0005] The purpose of this invention is to provide a method for constructing a three-dimensional model of gift box packaging design, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for constructing a three-dimensional model of a gift box packaging design, the method comprising:

[0007] Collect diverse design input data submitted by users, including natural language descriptions, hand-drawn sketches, and physical material property parameters;

[0008] Perform intent parsing on natural language descriptions, extract key design terms, and construct a design intent graph;

[0009] Perform structured feature analysis on hand-drawn sketch images to identify outline elements and decorative patterns within the sketches and generate geometric feature codes;

[0010] By aligning and fusing design intent maps with geometric feature encodings to form an enhanced design semantic representation;

[0011] Generative design algorithms are used to generate three-dimensional concepts from the enhanced design semantic representation, producing an initial three-dimensional design framework;

[0012] Integrate physical material property parameters into the initial three-dimensional design framework, perform material compatibility simulation verification, and obtain a feasibility assessment report;

[0013] Based on the feasibility assessment report, an adaptive parameter adjustment cycle is initiated to dynamically optimize the component parameters of the 3D design framework;

[0014] The optimized 3D model is visualized through a real-time rendering engine, and the final design documents are delivered.

[0015] Preferably, the step of performing intent parsing processing on the natural language description, extracting key design terms, and constructing a design intent map includes:

[0016] Use syntax analysis tools to decompose the sentence structure in natural language descriptions and identify the subject, attribute modifiers, and action instructions;

[0017] From the decomposition results, professional terms related to gift box design were selected to form a candidate terminology set;

[0018] The domain knowledge base is queried to verify the validity of candidate terms, remove ambiguous terms, and supplement synonym expansion;

[0019] A node relationship network is established based on the validated terminology set, where nodes represent design elements and edges represent spatial or logical relationships between elements.

[0020] The graph embedding algorithm is used to encode the node relationship network into a low-dimensional vector to generate a design intent graph.

[0021] Preferably, the step of performing structured feature analysis on the hand-drawn sketch image, identifying contour elements and decorative patterns within the sketch, and generating geometric feature codes includes:

[0022] Preprocessing operations are performed on the hand-drawn sketch image, including line enhancement and noise removal, to obtain a clear sketch;

[0023] An edge detection algorithm is used to extract continuous contour lines from the sketch and divide the contour lines into independent geometric units;

[0024] Calculate a shape descriptor for each geometric unit, including curvature distribution and symmetry index;

[0025] Identify repeating patterns and texture regions within geometric units, extract local feature points, and generate a feature description matrix;

[0026] The shape descriptors and feature description matrices of all geometric units are aggregated into a hierarchical feature structure, and the geometric feature codes are output.

[0027] Preferably, the step of knowledge-aligning and fusing the design intent map with geometric feature encoding to form an enhanced design semantic representation includes:

[0028] The vector representation of the design intent map is projected onto the vector space of the geometric feature encoding, and the semantic similarity matrix between the two is calculated.

[0029] A mapping table between intents and geometric elements is established based on the semantic similarity matrix;

[0030] By adjusting the weights of geometric feature encoding through a cross-modal attention mechanism, feature dimensions that are highly relevant to the design intent are highlighted;

[0031] The weighted geometric feature codes are concatenated with the design intent map, and then dimensionality is reduced through a nonlinear transformation layer to generate an enhanced design semantic representation.

[0032] Preferably, the step of using a generative design algorithm to generate a three-dimensional concept from the enhanced design semantic representation, producing an initial three-dimensional design framework, includes:

[0033] The design will enhance the semantic representation input of the pre-trained 3D generative neural network, which consists of an encoder and a decoder;

[0034] The encoder converts the enhanced design semantic representation into latent space vectors, the decoder generates a voxel mesh based on the latent space vectors, performs surface reconstruction on the voxel mesh, and generates a smooth 3D mesh model.

[0035] Key structural points are extracted from the 3D mesh model, a parametric skeleton system is constructed, and the parametric skeleton system is bound to the 3D mesh model to form an editable initial 3D design framework.

[0036] Preferably, the process of integrating physical material property parameters into the initial three-dimensional design framework, performing material compatibility simulation verification, and obtaining a feasibility assessment report includes:

[0037] Analyze the elastic modulus, density, and surface friction coefficient among the physical material properties;

[0038] Set the material parameters in the physics engine and import the initial 3D design framework into the simulation environment;

[0039] Virtual external forces are applied to simulate the load-bearing and deformation behavior of gift boxes, and stress distribution data is recorded;

[0040] Check whether stress concentration areas exceed the material strength threshold and check whether interference occurs between components;

[0041] Generate a detailed report, including stress exceedance points and interference issues, as a feasibility assessment report.

[0042] Preferably, the step of initiating an adaptive adjustment cycle of parameters based on the feasibility assessment report to dynamically optimize the component parameters of the 3D design framework includes:

[0043] Extract the location and severity indicators of the problem area from the feasibility assessment report. Based on the problem area, locate the corresponding component parameters, including thickness, angle and connection method, and establish a parameter optimization objective function. With the constraints of minimizing stress and avoiding interference, use the gradient descent algorithm to iteratively adjust the component parameters. After each iteration, rerun the material compatibility simulation test. When the improvement rate of three consecutive iterations is lower than the threshold, terminate the loop and output the optimized three-dimensional design framework.

[0044] Preferably, the 3D model visualized and optimized via a real-time rendering engine includes:

[0045] The optimized 3D design framework is converted into a rendering-friendly mesh format, and virtual lighting environment and material maps are set to make the model present a realistic visual effect.

[0046] It allows users to interactively rotate and zoom the model, view details from different perspectives, generate dynamic preview animations, demonstrate the opening and closing process of the gift box, and output high-resolution rendered images and video files as the final design files.

[0047] Preferably, the method further includes designing an iterative learning mechanism:

[0048] Record the parameter modification history and simulation results in each parameter adaptive adjustment cycle, build a design decision database, store successful and unsuccessful design cases, and use machine learning algorithms to analyze the design decision database to discover the implicit patterns between parameter adjustment and simulation results;

[0049] When dealing with new design tasks, prioritize applying the learned patterns to recommend initial parameter settings to accelerate convergence.

[0050] Preferably, the method further includes a multi-user collaborative optimization process:

[0051] Multiple designers are allowed to submit design input data simultaneously and merge it into a collective design intent. A voting mechanism is used to select the most popular geometric feature encodings to form a consensus design semantic representation. A conflict detection algorithm is introduced in the parameter adaptive adjustment loop to resolve parameter preference conflicts between different designers. Finally, after the 3D model is generated, feedback scores from each designer are collected to update the collaborative optimization strategy.

[0052] Compared with the prior art, the beneficial effects of the present invention are:

[0053] By constructing a design intent map through intent parsing of natural language and generating geometric feature codes through structured feature analysis of sketch images, the two are aligned and fused to form an enhanced design semantic representation. This process achieves the mapping and unification of the text semantic space and the sketch geometric space, enabling stylistic terms to establish precise correspondences with specific contour lines and decorative patterns identified in the sketch. Multimodal design input information is no longer isolated data fragments, but constitutes an organic whole with internal connections and mutual interpretations. This enhances the depth of understanding and the accuracy of representation of the designer's complex and ambiguous creative intent, ensuring from the source that the generated 3D conceptual framework faithfully reflects the original idea.

[0054] The physical material properties are integrated into the initial 3D design framework, and material compatibility simulations are performed. Based on the resulting feasibility assessment report, an adaptive parameter adjustment cycle is initiated. Material constraints are pre-emptively incorporated into the design optimization process, and the system automatically adjusts component parameters such as box wall thickness and snap-fit ​​structure gaps based on simulation results. This mechanism transforms the previously reactive, passive material verification into an active optimization-driven process throughout the design phase, ensuring that the 3D model continuously meets the manufacturability requirements of the physical world during its evolution. Design iteration shifts from an external cycle dependent on manual intervention to an endogenous adaptive cycle, reducing the risk of rework due to material and structural mismatches and shortening the cycle from concept to a productionable solution. Attached Figure Description

[0055] Figure 1 This is a schematic diagram illustrating the working principle of the method for constructing a three-dimensional model of gift box packaging design according to the present invention.

[0056] Figure 2 A flowchart for constructing a design intent map;

[0057] Figure 3 Flowcharts generated to enhance the semantic representation of the design;

[0058] Figure 4 A graph showing the relationship between the objective function and the constraints;

[0059] Figure 5 Bar chart showing the weight distribution of importance parameters for gift box packaging design. Detailed Implementation

[0060] 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.

[0061] Please see Figure 1 This invention provides a method for constructing a 3D model of gift box packaging design. The method includes: collecting diverse design input data submitted by users, including natural language descriptions, hand-drawn sketches, and physical material property parameters. The natural language descriptions are processed to extract key design terms and construct a design intent map. The hand-drawn sketches are analyzed using structured features to identify contour elements and decorative patterns and generate geometric feature codes. The design intent map and geometric feature codes are fused using knowledge alignment to form an enhanced design semantic representation. A generative design algorithm is used to generate 3D concepts from the enhanced design semantic representation, producing an initial 3D design framework. Physical material property parameters are integrated into the initial 3D design framework, and material compatibility simulation is performed to obtain a feasibility assessment report. Based on the feasibility assessment report, a parameter adaptive adjustment loop is initiated to dynamically optimize the component parameters of the 3D design framework. The optimized 3D model is visualized using a real-time rendering engine, and the final design file is delivered.

[0062] Example 1: See Figure 2In a specific implementation, the intent parsing process for natural language descriptions involves multiple steps. A grammatical analysis tool is used to decompose the sentence structure in the natural language description, identifying the subject, attribute modifiers, and action instructions. From the decomposition results, professional terms related to gift box design are selected to form a candidate term set. The validity of the candidate terms is verified by querying a domain knowledge base, ambiguous terms are removed, and synonym expansion is added. A node relationship network is built based on the verified term set, where nodes represent design elements, and edges represent spatial or logical relationships between elements. A graph embedding algorithm is applied to encode the node relationship network into a low-dimensional vector, generating a design intent graph. In a specific implementation, the grammatical analysis tool uses a dependency parsing method to identify the subject-verb-object structure and modifiers in the sentence, thereby extracting entities and attributes related to gift box design. In some embodiments, the domain knowledge base contains a professional vocabulary list and historical design case data in the gift box design domain to verify the accuracy and consistency of the terms. Optionally, synonym expansion is based on semantic similarity calculation, retrieving similar terms from the knowledge base to enrich the candidate set. It can be understood that the construction of the node relationship network relies on a graph data structure, where the edge weights between nodes reflect the strength of element associations. In practical implementation, the graph embedding algorithm uses the following formula to calculate the node embedding vector:

[0063] ;

[0064] in: Represents a node Embedded vector, It is a node The set of neighboring nodes, It is a trainable weight matrix. It is a node eigenvectors, It is a bias term. In some embodiments, the feature vector Semantic encoding derived from terminology, obtained through a pre-trained language model. Optional, a set of neighbor nodes. Based on the spatial or logical relationships between nodes, such as containment or adjacency relationships, the dimension of the embedding vector is set via hyperparameters to balance expressive power and computational efficiency. In practice, the generated design intent graph serves as a low-dimensional vector representation for subsequent knowledge alignment and fusion processes.

[0065] Example 2: See Figure 3In a specific implementation, performing structured feature analysis on a hand-drawn sketch image involves preprocessing operations, including line enhancement and noise removal to obtain a clear sketch. An edge detection algorithm is used to extract continuous contour lines from the sketch and segment these contour lines into independent geometric units. For each geometric unit, a shape descriptor is calculated, including curvature distribution and symmetry indices. Repeating patterns and texture regions within the geometric unit are identified, and local feature points are extracted to generate a feature description matrix. The shape descriptors and feature description matrices of all geometric units are aggregated into a hierarchical feature structure, and geometric feature codes are output. In a specific implementation, line enhancement uses morphological operations to enhance the continuity of sketch lines, noise removal uses filtering algorithms to reduce image imperfections, and the edge detection algorithm uses the Canny operator to identify contour boundaries and segment contour lines into independent geometric units based on connection points. In some embodiments, shape descriptor calculation includes analyzing the curvature changes at geometric unit boundary points using a curvature scale space method, and the symmetry index evaluates the symmetry degree of the geometric units using a mirror symmetry detection algorithm. Optionally, local feature point extraction uses the SIFT algorithm to generate the feature description matrix, and the hierarchical feature structure integrates the shape descriptors and feature description matrices using a tree-like data organization method. The specific implementation of the local feature point extraction using the SIFT algorithm to generate a feature description matrix includes the following steps. First, after preprocessing the sketch image to obtain a clear sketch, the system identifies repeating patterns and texture regions within geometric units. Then, the SIFT algorithm is applied to detect key points in these regions as local feature points, and a feature description matrix is ​​generated based on the surrounding image information of the key points. This matrix is ​​used to quantify the unique attributes of the local features. Next, the system integrates the shape descriptors (such as curvature distribution and symmetry indices) of each geometric unit with the feature description matrix using a tree-like data organization. The tree structure hierarchically associates the feature data of different geometric units, ultimately forming a unified geometric feature encoding output.

[0066] In practical implementation, when aligning and fusing the design intent map and geometric feature encoding, the vector representation of the design intent map is projected onto the vector space of the geometric feature encoding, and a semantic similarity matrix is ​​calculated between the two. Based on the semantic similarity matrix, a mapping table between intent and geometric elements is established. A cross-modal attention mechanism is used to adjust the weights of the geometric feature encoding to highlight features highly relevant to the design intent. The weighted geometric feature encoding is then concatenated with the design intent map and dimensionality is reduced using a nonlinear transformation layer to generate an enhanced design semantic representation. In some embodiments, the projection operation uses a linear transformation matrix to map the design intent map vector to the vector space of the geometric feature encoding. The formula for calculating the semantic similarity matrix is: ;

[0067] in: Represents the semantic similarity matrix. This represents the projected design intent vector. Represents the geometric feature encoding vector. This represents the matrix transpose. Optionally, the mapping table generates the correspondence between intent and geometric elements based on a threshold filtering of the semantic similarity matrix, and the cross-modal attention mechanism uses a query-key-value attention model to compute the weight vector.

[0068] Example 3: In a specific implementation, when using a generative design algorithm to generate 3D concepts from the enhanced design semantic representation, the enhanced design semantic representation is input into a pre-trained 3D generative neural network. This neural network consists of an encoder and a decoder. The encoder converts the enhanced design semantic representation into latent space vectors, and the decoder generates a voxel mesh based on these latent space vectors. The voxel mesh undergoes surface reconstruction to generate a smooth 3D mesh model. Key structural points are extracted from the 3D mesh model to construct a parametric skeleton system. The parametric skeleton system is then bound to the 3D mesh model to form an editable initial 3D design framework. In this specific implementation, the 3D generative neural network employs a variational autoencoder architecture. The encoder consists of fully connected layers and convolutional layers, while the decoder consists of deconvolutional layers and upsampling layers. The latent space vectors are calculated through the forward propagation of the encoder. In some embodiments, the voxel mesh generation process is represented by the following formula: ;

[0069] in: This represents the generated voxel mesh. Indicates decoder, Indicates encoder, This enhances the semantic representation of the design. Optionally, the surface reconstruction process uses a moving cube algorithm to extract isosurfaces from the voxel mesh, and key structural points are determined through mesh curvature analysis and feature point detection algorithms. In a specific implementation, mesh curvature analysis involves calculating the curvature values ​​of each point on the surface of the 3D mesh model, identifying high-curvature regions by analyzing the curvature distribution. These regions typically correspond to edges or corners and serve as key structural points. In another implementation, the feature point detection algorithm is applied to the 3D mesh to detect local geometric features such as corner points or texture change points, determining the location of feature points by comparing the geometric properties of neighboring points. The algorithm scans the mesh surface, evaluates the local geometric changes at each point, and uses a feature response function to filter salient points. These points are used to identify important structural points and support the construction of a parametric skeleton system. It can be understood that the parametric skeleton system consists of joints and connecting bones, and the binding operation associates the vertices of the 3D mesh model with the skeleton system through skinning weights.

[0070] In practical implementation, when integrating physical material property parameters into the initial 3D design framework to perform material compatibility simulation verification, the elastic modulus, density, and surface friction coefficient in the physical material property parameters are analyzed. Material parameters are set in the physics engine, and the initial 3D design framework is imported into the simulation environment. Virtual external forces are applied to simulate the load-bearing and deformation behavior of the gift box, and stress distribution data is recorded. It is then used to detect whether stress concentration areas exceed the material strength threshold and to check for interference between components. A detailed report containing stress exceedance points and interference issues is generated as a feasibility assessment report. In some embodiments, the physics engine employs a simulation system based on the finite element method, and the material parameters are configured according to the physical material property parameters, including elastic modulus, density, and surface friction coefficient. Optionally, virtual external forces include pressure loads and torque loads, and stress distribution data is obtained by solving mechanical equilibrium equations. It can be understood that stress concentration area detection is achieved by comparing nodal stress values ​​with the material strength threshold, and interference checks between components are performed by identifying mesh penetration phenomena through collision detection algorithms. The feasibility assessment report records all detection results in a structured document format.

[0071] Example 4: In a specific implementation, an adaptive parameter adjustment loop is initiated based on the feasibility assessment report to dynamically optimize the component parameters of the 3D design framework. The location and severity indicators of the problem area are extracted from the feasibility assessment report. Based on the component parameters corresponding to the problem area location, including thickness, angle, and connection method, a parameter optimization objective function is established with the constraints of minimizing stress and avoiding interference. The gradient descent algorithm is used to iteratively adjust the component parameters. After each iteration, a material compatibility simulation is run again. When the improvement rate of three consecutive iterations is lower than a threshold, the loop terminates and the optimized 3D design framework is output. In this implementation, the location of the problem area is identified using 3D coordinates, the severity indicator is quantified based on stress exceedance values ​​and interference distance, and the component parameters thickness, angle, and connection method are extracted from the metadata of the 3D design framework. The parameter optimization objective function is defined as: ;

[0072] in: Represents the objective function value. The component parameter vector represents the thickness, angle, and connection method. and These are weighting coefficients. Indicates the first Stress values ​​at each node, Indicates the first The distance value of each interference point, and These represent the number of nodes and interference points, respectively. In some embodiments, the gradient descent algorithm calculates the partial derivative of the objective function with respect to the parameter vector to update the parameters. After each iteration, the material compatibility simulation test is rerun to obtain new stress distribution data and interference detection results. Optionally, the improvement rate is calculated by comparing the changes in the objective function value in consecutive iterations, with a threshold set as a preset percentage. It is understood that the termination loop condition ensures that the optimization process stops when it converges and stabilizes. In a specific implementation, referring to Table 1, the parameter adaptive adjustment loop involves the mapping relationship between the problem region and component parameters.

[0073] Table 1: Mapping Table of Problem Areas and Component Parameters

[0074]

[0075] In some embodiments, component thickness is adjusted via linear transformation, angle via rotational transformation, and connection method via enumeration selection. Optionally, material compatibility simulation checks use the same physics engine settings in each iteration to ensure consistency. It is understood that the optimized 3D design framework is output in a standard mesh format.

[0076] In practical implementation, the optimized 3D model is visualized through a real-time rendering engine. The optimized 3D design framework is converted into a rendering-friendly mesh format. A virtual lighting environment and material maps are set to give the model a realistic visual effect. Users can interactively rotate and zoom the model and view details from different perspectives. A dynamic preview animation is generated to show the opening and closing process of the gift box. High-resolution rendered images and video files are output as the final design files. In practical implementation, the rendering-friendly mesh format includes OBJ or FBX format. The virtual lighting environment uses physically based rendering technology to set point light sources and directional light sources. In some embodiments, diffuse maps and normal maps are applied according to physical material property parameters. User interaction is achieved through a graphical user interface for rotation and scaling operations. Optionally, the dynamic preview animation simulates the hinge movement of the gift box using keyframe animation technology. High-resolution rendered images and video files are output in PNG and MP4 formats. It is understood that the final design file contains metadata describing design parameters and material properties.

[0077] See Figure 4The figures visually illustrate the relationship between the objective function value, average stress value, and average interference distance, providing quantitative guidance for the parameter optimization process. In Figure a, the relationship between the objective function value and the average stress value is shown. The horizontal axis represents the average stress value (MPa), the vertical axis represents the objective function value, the dashed line represents the stress threshold (60MPa), and the color bar on the right represents the number of iterations. As the average stress value gradually increases from 40MPa to 120MPa, the objective function value generally shows a decreasing trend. Even after the average stress value exceeds the stress threshold of 60MPa, the decreasing trend of the objective function value continues, and the number of iterations decreases as the average stress value increases. In Figure b, the relationship between the objective function value and the average interference distance is shown. The horizontal axis represents the average interference distance (mm), the vertical axis represents the objective function value, the dashed line represents the minimum interference distance (2mm), and the color bar on the right represents the number of iterations. As the average interference distance gradually increases from 0mm to 4.0mm, the objective function value generally shows an increasing trend. Even after the average interference distance exceeds the minimum interference distance of 2mm, the increasing trend of the objective function value continues, and the number of iterations increases as the average interference distance increases. These relationships provide a quantitative basis for iterative adjustment of component parameters (thickness, angle, connection method, etc.) based on gradient descent algorithm. By monitoring the changes in objective function value, average stress value, and average interference distance in real time, it is possible to determine whether parameter optimization has converged, ensuring that the stress and interference problems of the final 3D design framework meet the design requirements.

[0078] Example 5: In a specific implementation, an iterative learning mechanism is designed to record the parameter modification history and simulation results in each parameter adaptive adjustment cycle. A design decision database is constructed to store successful and failed design cases. Machine learning algorithms are used to analyze the design decision database to discover implicit patterns between parameter adjustments and simulation results. When processing new design tasks, the learned patterns are preferentially applied to recommend initial parameter settings to accelerate convergence. In a specific implementation, the parameter modification history includes adjustment records of thickness, angle, and connection method. The simulation results include stress distribution data and interference detection results. The design decision database uses a relational database structure to store the mapping relationship between parameter vectors and simulation results. In some embodiments, the machine learning algorithm uses a random forest regression model to analyze the relationship between parameter adjustments and simulation results. The implicit patterns are obtained through feature importance analysis to obtain the influence weight of parameters on simulation results. Optionally, the following optimization formula is recommended for initial parameter settings: ;

[0079] in: This represents the recommended initial parameter vector. Represents a parameter vector. Indicates the number of similar cases. Indicates the first The weighting coefficients of each case. This represents the prediction result corresponding to the parameter vector. Indicates the first The actual simulation results for this case demonstrate that faster convergence is achieved by reducing the number of iterations in the parameter adaptive adjustment loop.

[0080] In practical implementation, the multi-user collaborative optimization process allows multiple designers to simultaneously submit design input data, which is then merged into a collective design intent. A voting mechanism is used to select the most popular geometric feature codes to form a consensus design semantic representation. A conflict detection algorithm is introduced in the parameter adaptive adjustment loop to resolve parameter preference conflicts among different designers. After the final 3D model is generated, feedback scores from each designer are collected to update the collaborative optimization strategy. Specifically, the collective design intent is generated by weighted fusion of design input data from multiple designers, and the voting mechanism ranks and selects based on designers' scores of the geometric feature codes. In some embodiments, the conflict detection algorithm identifies mutually exclusive options in parameter settings and resolves conflicts using a negotiation strategy. Parameter preference conflicts include differences in thickness settings and disagreements on connection method selection. Optionally, feedback scores use a five-point scale to collect designers' evaluations of each dimension of the final 3D model, and the collaborative optimization strategy adjusts the weight allocation of the voting mechanism based on the feedback scores. It can be understood that updating the collaborative optimization strategy optimizes the weight allocation parameters through reinforcement learning algorithms, improving the efficiency of subsequent multi-user collaborative design.

[0081] See Figure 5 This paper presents the importance weight distribution of various physical material properties and structural parameters. Thickness (mm) has an importance weight of 0.32, making it a key parameter affecting the design; angle (degrees) has a weight of 0.25, connection method has a weight of 0.18, material density has a weight of 0.10, elastic modulus has a weight of 0.08, and surface friction coefficient has a weight of 0.07. These weights reflect the degree of influence of each parameter on simulation results (such as stress distribution and interferometry) in the design iterative learning mechanism. This provides an intuitive basis for parameter importance analysis based on machine learning algorithms and offers a quantitative reference for resolving parameter preference conflicts and recommending initial parameters in multi-user collaborative optimization processes, helping to accelerate design convergence and improve design efficiency.

[0082] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0083] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a three-dimensional model of a gift box packaging design, characterized in that, The method performs the following sequence of operations: Collect diverse design input data submitted by users, including natural language descriptions, hand-drawn sketches, and physical material property parameters; Perform intent parsing on natural language descriptions, extract key design terms, and construct a design intent graph; Perform structured feature analysis on hand-drawn sketch images to identify outline elements and decorative patterns within the sketches and generate geometric feature codes; By aligning and fusing design intent maps with geometric feature encodings to form an enhanced design semantic representation; Generative design algorithms are used to generate three-dimensional concepts from the enhanced design semantic representation, producing an initial three-dimensional design framework; Integrate physical material property parameters into the initial three-dimensional design framework, perform material compatibility simulation verification, and obtain a feasibility assessment report; Based on the feasibility assessment report, an adaptive parameter adjustment cycle is initiated to dynamically optimize the component parameters of the 3D design framework; Visualize the optimized 3D model using a real-time rendering engine and deliver the final design files; The step of initiating an adaptive adjustment cycle of parameters based on the feasibility assessment report to dynamically optimize the component parameters of the 3D design framework includes: Extract the location and severity indicators of the problem area from the feasibility assessment report. Based on the problem area, locate the corresponding component parameters, including thickness, angle and connection method, and establish a parameter optimization objective function. With the constraints of minimizing stress and avoiding interference, use the gradient descent algorithm to iteratively adjust the component parameters. After each iteration, rerun the material compatibility simulation test. When the improvement rate of three consecutive iterations is lower than the threshold, terminate the loop and output the optimized three-dimensional design framework.

2. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The process of performing intent parsing on the natural language description, extracting key design terms, and constructing a design intent map includes: Use syntax analysis tools to decompose the sentence structure in natural language descriptions and identify the subject, attribute modifiers, and action instructions; From the decomposition results, professional terms related to gift box design were selected to form a candidate terminology set; The domain knowledge base is queried to verify the validity of candidate terms, remove ambiguous terms, and supplement synonym expansion; A node relationship network is established based on the validated terminology set, where nodes represent design elements and edges represent spatial or logical relationships between elements. The graph embedding algorithm is used to encode the node relationship network into a low-dimensional vector to generate a design intent graph.

3. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The process of performing structured feature analysis on the hand-drawn sketch image, identifying contour elements and decorative patterns within the sketch, and generating geometric feature codes includes: Preprocessing operations are performed on the hand-drawn sketch image, including line enhancement and noise removal, to obtain a clear sketch; An edge detection algorithm is used to extract continuous contour lines from the sketch and divide the contour lines into independent geometric units; Calculate a shape descriptor for each geometric unit, including curvature distribution and symmetry index; Identify repeating patterns and texture regions within geometric units, extract local feature points, and generate a feature description matrix; The shape descriptors and feature description matrices of all geometric units are aggregated into a hierarchical feature structure, and the geometric feature codes are output.

4. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The process of aligning and fusing design intent maps with geometric feature encodings to form an enhanced design semantic representation includes: The vector representation of the design intent map is projected onto the vector space of the geometric feature encoding, and the semantic similarity matrix between the two is calculated. A mapping table between intents and geometric elements is established based on the semantic similarity matrix; By adjusting the weights of geometric feature encoding through a cross-modal attention mechanism, feature dimensions that are highly relevant to the design intent are highlighted; The weighted geometric feature codes are concatenated with the design intent map, and then dimensionality is reduced through a nonlinear transformation layer to generate an enhanced design semantic representation.

5. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The process employs a generative design algorithm to generate three-dimensional concepts from the enhanced design semantic representation, producing an initial three-dimensional design framework, including: The design will enhance the semantic representation input of the pre-trained 3D generative neural network, which consists of an encoder and a decoder; The encoder converts the enhanced design semantic representation into latent space vectors, the decoder generates a voxel mesh based on the latent space vectors, performs surface reconstruction on the voxel mesh, and generates a smooth 3D mesh model. Key structural points are extracted from the 3D mesh model, a parametric skeleton system is constructed, and the parametric skeleton system is bound to the 3D mesh model to form an editable initial 3D design framework.

6. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The physical material property parameters are integrated into the initial three-dimensional design framework, and material compatibility simulation is performed to obtain a feasibility assessment report, including: Analyze the elastic modulus, density, and surface friction coefficient among the physical material properties; Set the material parameters in the physics engine and import the initial 3D design framework into the simulation environment; Virtual external forces are applied to simulate the load-bearing and deformation behavior of gift boxes, and stress distribution data is recorded; Check whether stress concentration areas exceed the material strength threshold and check whether interference occurs between components; Generate a detailed report, including stress exceedance points and interference issues, as a feasibility assessment report.

7. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The optimized 3D model, visualized through a real-time rendering engine, includes: The optimized 3D design framework is converted into a rendering-friendly mesh format, and virtual lighting environment and material maps are set to make the model present a realistic visual effect. It allows users to interactively rotate and zoom the model, view details from different perspectives, generate dynamic preview animations, demonstrate the opening and closing process of the gift box, and output high-resolution rendered images and video files as the final design files.

8. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The method also includes designing an iterative learning mechanism: Record the parameter modification history and simulation results in each parameter adaptive adjustment cycle, build a design decision database, store successful and unsuccessful design cases, and use machine learning algorithms to analyze the design decision database to discover the implicit patterns between parameter adjustment and simulation results; When dealing with new design tasks, prioritize applying the learned patterns to recommend initial parameter settings to accelerate convergence.

9. The method for constructing a three-dimensional model of gift box packaging design according to claim 1, characterized in that, The method also includes a multi-user collaborative optimization process: Multiple designers are allowed to submit design input data simultaneously and merge it into a collective design intent. A voting mechanism is used to select the most popular geometric feature encodings to form a consensus design semantic representation. A conflict detection algorithm is introduced in the parameter adaptive adjustment loop to resolve parameter preference conflicts between different designers. Finally, after the 3D model is generated, feedback scores from each designer are collected to update the collaborative optimization strategy.

Images