Color type linkage design method for gift set style unification

Abstract

The invention provides a color type linkage design method for gift set style unification, and relates to the technical field of product design, a gift set design demand and a style intention are obtained, a style label set with a weight is constructed by using a style knowledge graph, and a color style vector and a modeling style vector are generated on the basis. And according to the style vector, establishing a color type linkage constraint model between the color and the modeling, including a color-modeling semantic linkage rule, cross-category style consistency constraint and process material feasibility limitation. And further performing joint optimization on color parameters and modeling parameters of each gift subitem through a multi-objective optimization solution engine to generate candidate design schemes, and performing quantitative evaluation and interactive optimization on the schemes by adopting a comprehensive style unified index formed by visual unification degree, color unification degree and modeling unification degree.

Description

Technical Field

[0001] This invention relates to the field of product design technology, specifically to a color and shape linkage design method for unified style in gift sets. Background Technology

[0002] Current gift set design typically relies on the designer's personal aesthetic experience, manually adjusting color schemes and proportions across different gift categories to achieve a so-called "series-based style unity." However, this traditional method suffers from several significant problems: First, the structural differences between different gift categories (such as gift boxes, bags, cards, and accessories) are substantial, lacking a unified style expression system, making it difficult to maintain consistency across multiple dimensions such as color, shape, and texture. Second, the linkage between color scheme and shape lacks calculable rules, requiring designers to constantly experiment to achieve subjectively satisfactory results, which is time-consuming and involves a large amount of repetitive work. Third, the existing design process lacks parametric connections with the production end, often requiring manual redrawing of production files, which easily leads to quality issues such as color differences, shape discrepancies, and layout misalignments. Therefore, there is an urgent need for a technical method that can unify the modeling of color style, shape style, semantic style, and technological feasibility, and automatically output a complete design scheme that meets the requirements of a unified style, thereby improving design efficiency, style consistency, and production feasibility. Summary of the Invention

[0003] Technical problems to be solved

[0004] To address the shortcomings of existing technologies, this invention provides a color-type linkage design method for unified style in gift sets, thus solving the problems of existing technologies.

[0005] Technical solution

[0006] To achieve the above objectives, the present invention provides the following technical solution: a color-type linkage design method for unifying the style of gift sets, the color-type linkage design method comprising the following steps:

[0007] Sp1. Obtain the design requirements and style intent of gift sets, collect the application scenarios, target audience, gift category composition and style keywords of the target gift set, and map the style keywords into a set of weighted style tags based on the pre-built style knowledge graph.

[0008] Sp2. Based on the set of style tags, construct corresponding color style vectors and shape style vectors. The color style vector is defined in a preset color space and is used to represent the hue, brightness, saturation and color matching mode of the main color, secondary color and embellishment color of the gift. The shape style vector is used to represent at least one of the curvature characteristics, aspect ratio, outline complexity, symmetry and texture density of the gift's appearance.

[0009] Sp3. Based on the color style vector and the shape style vector, establish a color-shape linkage constraint model between the color and shape of the gift. The constraint model includes color-shape linkage rules based on style semantics and cross-category style consistency constraints for each sub-item of the gift set.

[0010] Sp4. Using the color-type linkage design solution engine, under the premise of satisfying the color-type linkage constraint model, jointly optimize the color parameters and shape parameters of each gift item in the gift set to generate at least one candidate complete color-type design scheme.

[0011] Sp5. Conduct a unified style quantitative evaluation of each candidate complete set of color design schemes to obtain a comprehensive style unity index, which includes visual unity index, color unity index and shape unity index. Based on the comprehensive style unity index, screen or iteratively optimize the candidate complete set of color design schemes.

[0012] Sp6. Output at least one set of color and shape design schemes for a complete set of gifts that meet the preset uniformity threshold, and record the color parameters and shape parameters of each gift item in a parametric form.

[0013] Preferably, in Sp1, the construction of the style tag set includes:

[0014] Expand the nodes of style keywords in the style knowledge graph to obtain candidate style tags associated with festival type, cultural attributes, aesthetic preferences and target audience;

[0015] Based on a pre-trained style tag relevance model or interactive adjustments by designers, weights are assigned to the candidate style tags to form a normalized set of style tags.

[0016] The style tag set is used to drive the determination of the target area of ​​the color style vector and the styling style vector.

[0017] Preferably, in Sp2, the construction of the color style vector includes:

[0018] In the CIELab color space, the colors in the preset color library are encoded to form a set of color vectors containing primary colors, secondary colors, and accent colors;

[0019] Based on the color vector set, add brightness level relationship parameters, saturation distribution parameters, and color harmony mode parameters;

[0020] Based on the style tag set and the preset color-style mapping function, the overall style of the gift set is mapped to a color style target vector region, and corresponding feasible color vector sub-regions are assigned to different gift categories in the gift set.

[0021] Preferably, in Sp2, the construction of the styling style vector includes:

[0022] For different gift categories, preset shape parameter templates are provided. The shape parameter templates include at least one of the following: control point coordinates of the outer contour of the gift, corner radius, length-width-height ratio, number of contour polyline nodes, and texture distribution density.

[0023] The styling parameter template is mapped to a styling style vector, and the target value range and allowable offset range of each styling parameter are determined according to the style tag set, thereby obtaining the styling style target vector of each gift category.

[0024] Preferably, in Sp3, the color-pattern linkage constraint model includes:

[0025] Color-shape semantic linkage rules are used to constrain the range of variation in curvature features, outline complexity, and texture density of the gift's appearance under a specific main color or color scheme.

[0026] Cross-category style consistency constraint is used to limit the distance between the color style vector and the shape style vector of each gift item in the gift set and the style center vector of the gift set, respectively, to no more than a preset threshold, and different offset tolerances are set for the main gift and the auxiliary gift.

[0027] Process and material constraints are used to limit the feasible set of color and shape parameters within the range of printing process, material strength, and dimensional specifications.

[0028] Preferably, in Sp4, the color-type linkage design solution engine is implemented by constructing a multi-objective optimization function, the corresponding multi-objective optimization function including:

[0029] The sub-objective function representing style consistency is used to measure the degree to which the color style vector and shape style vector of each gift item in the gift set deviate from the style center vector;

[0030] A sub-objective function characterizing the degree to which color-shape linkage rules are satisfied, used to measure the proportion or degree to which color-shape semantic linkage rules are satisfied;

[0031] A sub-objective function characterizing aesthetic harmony is used to score a complete set of color scheme design schemes based on a pre-trained aesthetic evaluation model or color harmony formula.

[0032] Sub-objective functions characterizing cost and process feasibility;

[0033] The color-type linkage design solution engine uses one or any combination of genetic algorithm, simulated annealing algorithm or particle swarm optimization algorithm to solve the above multi-objective optimization function to obtain the candidate complete color-type design scheme.

[0034] Preferably, in Sp5, the unified style quantitative evaluation of candidate complete color scheme design schemes includes:

[0035] Input the 2D renderings or 3D renderings of each gift item in the gift set into a pre-trained visual style recognition network to extract visual style feature vectors;

[0036] The visual style feature vector is used to calculate the feature similarity distribution within the gift set to obtain a visual unity index.

[0037] The visual unity index is weighted and fused with the color unity index and the form unity index calculated based on the color style vector and the form style vector to form a comprehensive style unity index.

[0038] Preferably, the Sp5 further includes:

[0039] When the overall style consistency index is lower than the preset consistency threshold, an interactive optimization process is triggered, displaying key color or shape parameters that have a significant impact on consistency to the designer.

[0040] The system receives adjustment instructions from the designer for the key parameters and, based on the adjusted parameters, calls the color-type linkage design solving engine again for local optimization to generate new candidate complete color-type design schemes.

[0041] Preferably, the system corresponding to the color-pattern linkage design method includes the following components:

[0042] The style intent acquisition module is used to obtain the design requirements and style intent for gift sets and generate a set of style tags;

[0043] The style vector construction module is used to construct color style vectors and styling style vectors based on the style tag set;

[0044] The color-shape linkage constraint construction module is used to establish a color-shape linkage constraint model consisting of color-shape linkage rules and cross-category style consistency constraints;

[0045] The color and shape linkage design solution module is used to jointly optimize the color parameters and shape parameters of each gift item in the gift set under the color and shape linkage constraint model, and generate candidate complete color and shape design schemes.

[0046] The unified style evaluation module is used to quantitatively evaluate the unified style of candidate complete color scheme design schemes and output a comprehensive style unification index.

[0047] The solution output module is used to output a set of color and shape design schemes for gifts that meet the preset uniformity threshold, along with their corresponding parametric color parameters and shape parameters.

[0048] Preferably, the system corresponding to the color-shape linkage design method further includes a parametric design data generation module, used to convert the color parameters and shape parameters in the selected complete set of color-shape design schemes into a structured output file that can be called by downstream design or production software. The structured output file includes:

[0049] CIELab color coordinates of the primary, secondary, and accent colors of each gift item and their regional distribution data on the gift surface;

[0050] Parametric description of shape parameters such as contour control point coordinates, fillet radius, aspect ratio, texture density, and shape complexity;

[0051] The overall style center vector of the gift set, the unified style index, and the style version number for production tracking;

[0052] The structured output file can be automatically parsed by packaging layout software, 3D modeling software, or printing plate making system to realize the automated implementation and mass production of gift set color and pattern design.

[0053] Beneficial effects

[0054] This invention provides a color-and-pattern linkage design method for creating unified styles in gift sets. It offers the following advantages:

[0055] 1. This invention transforms the design of gift sets from relying on subjective aesthetic judgment to a calculable "style vector—linkage constraint—optimization solution" system. By constructing color style vectors, shape style vectors, and a color-shape semantic linkage model, the system enables the coordinated calculation of parameters such as the main color, secondary color, accent color, shape curvature, contour complexity, and texture density of the gift within a unified style space. This achieves global consistency in visual style across categories, structures, and materials. The system automatically outputs design schemes that meet the uniformity threshold, significantly reducing the cost of repeated manual parameter adjustments and substantially improving the style quality and consistency controllability of gift sets.

[0056] 2. This invention constructs a parametric design data generation module, outputting the final color parameters (CIELab values, area distribution) and shape parameters (contour control points, corner radii, proportional dimensions, texture density, etc.) as structured files. This allows typesetting software, 3D modeling systems, and printing plate-making systems to automatically parse these design parameters, thereby achieving a highly automated closed loop from "design – proofing – production." This output mechanism not only avoids errors caused by manual secondary data entry but also makes design versions traceable and processes verifiable, truly realizing the standardization, visualization, and mass production capabilities of the gift design process. Attached Figure Description

[0057] Figure 1 This is a diagram illustrating the overall architecture of the present invention;

[0058] Figure 2 This is a flowchart of the method of the present invention;

[0059] Figure 3 This is a screenshot of the main system of the present invention;

[0060] Figure 4 This is a screenshot of the system style vector of the present invention;

[0061] Figure 5 This is a partial screenshot of the system design scheme of the present invention;

[0062] Figure 6 The screenshot shows a candidate system solution for this invention. Detailed Implementation

[0063] 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. Specific Implementation Example 1:

[0065] like Figures 1 to 6As shown, a color-type linkage design method for unified style in gift sets is presented. In the Sp1 gift set design requirements and style intent acquisition stage, the system first constructs a standardized design task data structure through a front-end interactive interface. This data structure can be in the form of a JSON object or a database record. Fields include at least an application scenario field, a target audience field, a gift category composition field, and a style keyword field. The application scenario field describes the type of festival, business purpose, wedding scene, or regional cultural background. The target audience field records information such as age group, gender ratio, consumption level, and aesthetic preference tags. The gift category composition field records the names, quantities, and relationships of categories such as main gifts, secondary gifts, packaging materials, and accessories. The style keyword field records the designer's subjective style intent in the form of natural language phrases or tags. The system internally stores a style knowledge graph, which can be implemented based on a graph database. Each node corresponds to a style concept, cultural symbol, festival attribute, or typical color scheme / typical shape feature, representing the semantic relationships and weights between concepts. The node attributes in the graph include style classification, related color features, related shape features, and common application scenarios. After receiving style keywords, the system calls the style keyword expansion function. This function performs a neighborhood search in the knowledge graph starting from the keywords, retrieving candidate style tags highly related to festival type, cultural attributes, aesthetic preferences, and target audience within a limited number of layers using a depth-first or breadth-first strategy. Subsequently, the system calls the style tag relevance calculation function. This function can calculate the matching degree between candidate style tags and the overall context of the current design task based on a vector representation model (e.g., embedding style tags into a style semantic vector space) or a classification model based on statistical learning, and normalize the calculation results to form a set of style tags with weights. In actual implementation, the relevance calculation function can take the following form: First, the application scenario, target audience, and gift category are encoded together into a task context vector. Then, each candidate style tag is encoded into a tag vector. The similarity score is output using a cosine similarity function or a multilayer perceptron, and normalized to a probabilistic weight using a softmax function. Finally, a set of style tags and corresponding weights are obtained. This set of style tags will serve as the input parameters for the subsequent color style vector and styling style vector target region determination function.

[0066] In the implementation of the style tag set construction and weight calculation described in Sp1, the context of the entire design task can be encoded into a vector form. Let the design task context be... ,in Integrating information such as application scenarios, target audience, and gift category composition, each style tag node in the style knowledge graph is denoted as... The corresponding semantic vector is The task context vector is The system first uses a similarity function. Calculate the matching degree between the context and each candidate style tag. The similarity can be expressed as a dot product or cosine similarity. For example, when using the dot product form, it can be written as...

[0067]

[0068] Then, the similarity scores are normalized into weights using the softmax function, and the style tag weights are denoted as... Then there is in For the set of candidate style tags participating in the calculation, all All are non-negative real numbers and satisfy In actual implementation, With e The acquisition of these weights can rely on pre-trained embedding models or vectorized networks fine-tuned with project data. This is the mathematical form of Sp1, which assigns weights to candidate style tags and forms a normalized style tag set based on a pre-trained style tag relevance model.

[0069] In the specific implementation of color style vector construction in SP2, the system pre-establishes a color library during the initial deployment phase, encoding commonly used printing colors, spot colors, and typical style colors into the CIELab color space. Each color is stored as a vector containing L, a, and b components, and associated with the feasibility information of the color in printing and materials, such as whether it is suitable for specific materials, whether it can be approximated by four-color printing, and the deviation range of digital proofing. When constructing the color style vector, the system needs to select several color combinations from the color library to form primary, secondary, and accent color sets, and perform a structured description of these color sets. Therefore, the color style vector is defined as a composite vector containing multiple sub-vectors. This composite vector includes not only the Lab coordinates of the primary, secondary, and accent colors, but also parameters such as the difference in brightness between the primary and secondary colors, the overall average brightness, saturation distribution statistics, the proportion of each color in the visual area, and the color matching mode encoding (e.g., complementary, analogous, three-color matching, etc.). When the system implements the color-style mapping function, it will... The system pre-learns or sets a set of target color regions for each style tag. For example, the target color region corresponding to the "New Chinese Style" tag can be defined as a combination of certain regions in the Lab space, such as a dark base with low brightness and low to medium saturation plus a high brightness, warm-toned accent color. The system generates the overall color style target vector region of the gift set by weighted synthesis of the target color regions of each tag in the style tag set. On this basis, the system also needs to divide the overall region into feasible color vector sub-regions according to the gift category. This process is achieved through a category allocation function. This function considers the importance weight of the main gift and the auxiliary gift, their visual dominance in the set, and human-computer experience rules, and appropriately shifts or shrinks the style center color region to leave suitable color space for different categories. For example, the main gift box usually occupies a large area in the image, so its main color needs to be closer to the overall style center, while the accessories can use slightly different hues or brightness within the allowable range to improve the sense of layering.

[0070] In the implementation of styling style vector construction in Sp2, the system first establishes a styling parameter template library for different gift categories. Each template describes a typical structural form, such as a cuboid gift box, a drawer-type gift box, cylindrical packaging, a tote bag, a card, etc. Each template is defined by a set of geometric parameters and morphological complexity parameters. Geometric parameters may include length, width, and height values, length-width-height ratio, corner radius, coordinates of contour control points, and whether it includes internal structures such as partitions or inner trays. Morphological complexity parameters can be quantified by counting the number of contour line segments, the number of inflection points of polylines, the number of curve control points, and the number of local details. Symmetry parameters can be obtained by calculating the overlap index after mirroring the outer contour shape along the main axis of symmetry. Texture density can be estimated by the number of texture units per unit area, texture repetition period, and texture scale distribution. The system encodes the aforementioned geometric and complexity information into a styling style vector. Each dimension corresponds to a continuous or discrete morphological index, and the system derives the target value range for these indices based on the weights of the style tag set. For example, when the "minimalist" style tag has a high weight, the system uses a styling-style mapping rule function to set the outline complexity target value to a lower range, the corner radius target value to a medium-to-high range, and the texture density target value to a low or even very low value. When the "tech" or "futuristic" tags have a high weight, the mapping rule can derive higher outline complexity, higher texture density, and target values ​​for local curvature of some sharp transitions. The styling style target vector ultimately provides a central vector and offset range for each gift category. The offset range can be expressed in the form of standard deviation or upper and lower bounds. In subsequent optimization processes, all adjustments to styling parameters must remain within this range or be penalized.

[0071] In the steps of Sp2 regarding the construction of color style vectors, each color can be represented as a three-dimensional vector in CIELab space, assuming the primary color is... Secondary color is The accent color is In addition, the proportion coefficients of the main color, secondary color, and accent color in the visual area are introduced. ,satisfy And define the overall average brightness. The mean value of lightness, a parameter related to saturation, can be expressed as: Saturation can be calculated using CIELab by considering the component beyond the projection length onto the L-axis. For example, the "chromaticity modulus" of each color can be defined as... The overall average saturation can then be written as: Based on this, a color style vector can be defined. for ,in Encoding color harmony patterns allows mapping different color relationships (such as complementary, analogous, and triadic) to integer codes or one-hot vectors. The description of "style tag set and preset colors—style mapping function maps the overall style to a color target vector region" can be formalized as: for each style tag... Predefine a set of color style prototype vectors and its covariance matrix The system obtains the weights based on Sp1. Calculate the overall target color center vector , At the same time, the target region can be defined as satisfying The ellipsoidal region in which It can be The weighted combination or empirically defined covariance matrix, A threshold is used to control the size of the region. When assigning color vector sub-regions to different gift categories, category weights can be introduced. , for the first Each product category defines its color center vector as follows: ,in For category-specific offset vectors, their upper bounds can be set according to the empirical rule that "main gifts have smaller offsets, secondary gifts have moderate offsets, and accessories have slightly larger offsets." .

[0072] In the construction of the styling vector in Sp2, the outer contour parameters and morphological complexity index of a certain gift sub-item can be uniformly represented as a vector. Assume the outer contour of the gift is defined by a set of control points. Description: Each control point is represented by a planar or spatial coordinate system. or The number of outline segments is denoted as The number of inflection points of a broken line or curve is denoted as Then the contour complexity index can be defined as ,in To control for the non-negativity of the weights of different constituent factors, the symmetry index can be measured by the degree of overlap between the contour and its mirror image along a principal axis. For example, let the set of contour points be... The set of points mirrored along the vertical axis is Then the symmetry score can be defined as ,in The number of sampling points involved in the calculation. The maximum reference distance used for normalization, therefore The value ranges roughly between 0 and 1, with larger values ​​indicating greater symmetry. Texture density can be estimated by the number of texture units per unit area. Let the surface area of ​​the gift be... The number of texture units is Then the texture density can be written as If the corner radius is also considered in the design and length-width-height ratio Then, a constructive style vector can be created. And also for each style tag Preset Prototype Vector With allowable variance ,pass weight Obtain the overall target styling center and through conditions To limit the style and design to a reasonable area.

[0073] When constructing the Sp3 color-style linkage constraint model, the system incorporates color style vectors and styling style vectors into a unified constraint system. This system consists of multiple constraint functions, each defining the relationship between color parameters and styling parameters, or across gift sub-items. At the color-styling semantic linkage level, the system defines a set of color-styling linkage rules. For example, when the main color is of low brightness and low saturation and points to the semantic nodes of "calm" and "elegant" in the style knowledge graph, the linkage rule function requires that the contour complexity dimension in the styling style vector does not exceed a medium value, and the proportion of sharp corners in the curvature feature must be below a certain threshold. Texture density must not be too high. This linkage rule can be formalized as a set of inequalities that constrain the shape vector components within a certain numerical range. For example, when the accent color has high saturation and strong color contrast, the linkage rule function requires that the visible area corresponding to the accent color must be mapped to a position with high curvature or local structural features, such as the border of the logo area, the handle connection, the edge of the open structure, etc. This can be achieved by establishing a color area mapping function, which maps color parameters to geometric area labels one by one and limits high saturation colors to appear only in shape areas marked as "small area high attention area". The cross-category style consistency constraint is achieved by defining a style center vector for the gift set. This style center vector is obtained by weighting the color style vector and shape style vector of all gift items according to a pre-set weight. For example, the main gift is given a larger weight, while the secondary gifts and accessories are given a smaller weight. Then, the consistency constraint function limits the Euclidean or cosine distance between each item's vector and the style center vector to no more than a certain threshold. The threshold can be configured differently according to different category roles. For example, the main gift is allowed to deviate from the style center very little to ensure that it becomes the visual style benchmark, while the secondary gifts are allowed to deviate more widely to ensure the variation and hierarchy of the set within the unity. The process and material constraint layer establishes constraint functions based on a preset color printing process database and material performance database. This database records the process feasibility, cost factor, and color difference range of each material and each color combination. The constraint function directly marks certain color parameters and incompatible material combinations as infeasible solutions. At the same time, it penalizes or excludes color schemes that exceed the printing press's color limit and applies infeasible marks to structural combinations that result in insufficient material compressive strength or insufficient packaging strength, ensuring that the final optimized solution is not only stylistically consistent but also production-ready.

[0074] exist In the color-type linkage constraint model, the first... The color style vector and the shape style vector of each gift item are concatenated to form the overall style vector. The overall style center vector of the gift set can be obtained through a weighted average, where the role weight of each sub-item is denoted as . ,satisfy ,but Cross-category style consistency constraints can be formalized as imposing constraints on each sub-item. in and These are the overall color and the central vector of the design, respectively. , The offset tolerance is set for different product categories. The color-shape semantic linkage rule can be incorporated into the optimization objective in the form of a penalty function. For example, when the style tag contains "elegant" and the main color has low brightness and moderate saturation, the outline complexity can be required not to exceed a certain upper limit. When violated, a penalty function is applied. Constraints are applied; when the accent color chroma... When the value is very high, the corresponding visual area ratio can be required. Not exceeding the threshold Otherwise, increase the penalty. The process and material constraints can be determined by a set of logic functions to determine whether the color and material combinations are feasible. If a combination is not feasible, the solution is marked as infeasible or given a maximum penalty value. This allows for automatic exclusion during optimization. in It is a constant that is much larger than the normal target value.

[0075] In the Sp4 color-shape linkage design solution stage, the system constructs a multi-objective optimization function to jointly search the color and shape parameters of each sub-item of the gift set. This multi-objective function is composed of several weighted sub-objective functions. Among them, the style consistency sub-objective function measures the average deviation of the color style vector and shape style vector of all gift sub-items from the style center vector. Usually, the weighted sum of the squares of the Euclidean distances of all sub-item vectors to the center vector can be used as the loss term. The color-shape linkage rule satisfaction sub-objective function calculates the satisfaction status of each color-shape semantic linkage rule and defines a satisfaction index for each rule. Full satisfaction is given as full marks, partial satisfaction is given intermediate scores using linear or nonlinear functions, and serious violations are given a large penalty. The overall objective is represented by a weighted average. The aesthetic harmony sub-objective function can be implemented through a pre-trained aesthetic evaluation model. During training, this model uses a large number of excellent complete gift sets or packaging design samples for supervised learning. The input is a generated 2D rendering or illustration, and the output is a unified aesthetic score. In actual operation, the system renders each candidate design and inputs it into the model to obtain an aesthetic score. This score is then normalized and incorporated into the overall objective function. The cost and process feasibility sub-objective function estimates the expected number of printing colors, special process usage, material usage, manufacturing complexity, and expected production loss rate for each design based on a process database and a material database. This results in a comprehensive cost and risk score, penalizing designs with high costs and risks. The overall multi-objective optimization function can be expressed as a weighted sum F = w1·Fstyle + w2·Flink + w3·Faesthetic − w4·Fcost, where the weights can be set by the designer or automatically adjusted by the system based on historical projects. To solve the objective function, the system employs one or more combinations of genetic algorithms, simulated annealing algorithms, or particle swarm optimization algorithms. In the implementation of the genetic algorithm, the system first randomly generates a batch of initial solutions based on the value ranges of the color vector and the shape vector. Each solution is encoded as a chromosome containing the color parameters and shape parameters of all gift sub-items. Then, an individual with a higher objective function value is selected to enter the breeding stage through a selection operator. A crossover operator is used to exchange some color or shape genes between two parent schemes. A mutation operator is used to perform small-probability random perturbations within the local parameter range. At the same time, all constraint functions are checked in each iteration. Individuals that violate hard constraints are directly eliminated, and individuals that violate soft constraints have their objective function values ​​adjusted through a penalty term. The entire iterative process stops after the objective function converges or reaches a preset number of generations, and several candidate complete color and shape design schemes are output.

[0076] In the multi-objective optimization function of Sp4, each sub-objective can be given a clear formula. The style consistency sub-objective function can be measured by the comprehensive deviation of all gift sub-items, defined as follows: ,in and Control the relative weight of color and form in consistency. The degree to which the color-form linkage rule is satisfied can be reflected by the weighted sum of the above penalty items. ,in For the first The penalty function corresponding to each linkage rule. Assuming the importance weight of this rule, the aesthetic harmony sub-objective function can use negative aesthetic scores as the loss term. Let the aesthetic score obtained from the scheme through the pre-trained aesthetic model be... To facilitate consistency with other losses, it can be defined as follows: This ensures that the higher the aesthetic score, the smaller this item becomes, making the overall objective function easier to minimize. The cost and process feasibility sub-objective function can be minimized by estimating the total cost. and process risks Expressed by weighted sum, in , Weights are assigned to costs and risks. Combining these sub-objectives allows us to construct an overall optimization objective function. in The coefficients are configurable; the optimization process involves optimizing the parameters while ensuring all parameters satisfy the hard constraints. The process of minimization, within the framework of genetic algorithms, simulated annealing, or particle swarm optimization, involves an objective function. It is computed on each candidate solution to guide selection, crossover, and mutation operations.

[0077] In the SP5 unified style quantitative evaluation stage, the system needs to conduct objective uniformity assessment and interactive optimization of candidate complete color and pattern design schemes. First, in terms of visual uniformity assessment, the system generates corresponding two-dimensional effect images or three-dimensional rendering images of gift sub-items for each candidate scheme. Then, these images are input into a pre-trained visual style recognition network. The visual network can adopt a convolutional neural network structure or a visual Transformer structure. The first few layers are used to extract low-level texture and edge features, the middle layers are used to extract color block layout and shape contour features, and the high layers are used to extract overall style semantic features. The system reads the feature vector of the network at a certain middle or high level as the visual style feature vector, and calculates the pairwise cosine similarity of the feature vectors of different sub-items in the same set. On this basis, the visual uniformity index is constructed using statistical measures such as average similarity, minimum similarity, and variance. The more concentrated the feature vectors are and the higher the similarity is within the same set, the higher the visual uniformity index. The system then calculates color uniformity and shape uniformity using color style vectors and shape style vectors respectively. By statistically analyzing the distance between each sub-item vector and the style center vector, two uniformity indices are obtained. Finally, a comprehensive style uniformity index is obtained using the weighted fusion function Utotal=α·Uvis+β·Ucolor+γ·Ushape. The system presets one or more uniformity thresholds to distinguish between schemes at different levels, such as "qualified," "needs optimization," and "unqualified." When the overall style uniformity index of a candidate scheme falls below a certain threshold, the system automatically enters an interactive optimization mode. The interface highlights a set of parameters that have the greatest impact on uniformity. These parameters can be obtained through a uniformity sensitivity analysis function, which involves making small-scale perturbations to individual parameters near the candidate scheme and observing the change in the overall uniformity index to determine the sensitivity of that parameter. The system presents these key parameters to the designer in the form of visual sliders, input boxes, or graphical interactive controls. The designer can fine-tune the parameters according to their subjective aesthetic intentions. After receiving the adjustment instructions, the system uses the adjusted parameters as new constraints or initial points and calls the color-type linkage design solver engine again for local optimization. Local optimization can be achieved by narrowing the parameter search range or reducing the number of iterations, improving the uniformity index while ensuring response speed, ultimately resulting in a design scheme optimized through human-computer collaboration.

[0078] In the uniformity evaluation of Sp5, various uniformity indicators can also be given explicit formulas. Let the rendered image of the gift sub-item in a candidate scheme be... The feature extraction function of the visual style recognition network is denoted as... Then the first The visual feature vectors of each sub-item are For any two sub-items and It can calculate the cosine similarity of their visual features. The system can take the average similarity of all different sub-item pairs as a visual uniformity index. , in For the quantity of gift items. Color uniformity index can be based on... and The definition of distance, in For the maximum reference distance used for normalization, such that Roughly between 0 and 1, a higher value indicates greater color uniformity. A similar definition can be used for the shape uniformity index. in This is the normalization factor for the vector distance of the modeling. The overall style unification index can be given in a weighted form. in The weights are non-negative and usually satisfy the following conditions: ,thereby The numerical ranges are clear and easy to compare; when Below the set threshold At this point, the system considers the solution to lack uniformity and requires interactive optimization.

[0079] When identifying "key parameters with significant impact on uniformity" in Sp5, a simple uniformity sensitivity analysis formula can be introduced. Let the current comprehensive uniformity index of a candidate solution be... A certain parameter, such as the brightness of the main color of the main gift box. Slightly disturbed A new comprehensive unified index was subsequently obtained. Therefore, the uniformity sensitivity of this parameter can be approximately defined as: The larger the value, the greater the impact of a small change in the parameter on the uniform index. The system can calculate similar sensitivity indices for a group of candidate key parameters and sort them from high to low values. The top-ranked parameters are highlighted on the interface, thus providing designers with targeted adjustment suggestions.

[0080] In the Sp6 output stage and system implementation, this invention not only outputs visually appealing renderings or illustrations, but also records the entire design scheme in a structured, parametric format so that downstream design and production systems can directly consume this data. To this end, the system includes a parametric design data generation module. Upon receiving the final scheme determined after unified evaluation and interactive optimization, this module extracts the Lab coordinates, color proportions, corresponding geometric application area identifiers, and shape parameter vectors of the primary, secondary, and accent colors of all gift sub-items from its internal data structure. It then constructs a structured output file according to a predefined data format specification, which can be JSON. The output file can be either XML or a custom markup format compatible with existing CAD / packaging software. Within this file, the system assigns a unique identifier to each gift item and records geometric parameters such as length, width, height, corner radius, outline control point coordinate array, texture pattern reference identifier, and texture density parameters under the corresponding node. It also records color parameters such as the Lab coordinates of each color, application area number, and relative area ratio. Furthermore, the file records the overall style center vector of the gift set, various uniformity indicators and a comprehensive style uniformity index, as well as a style version number field. This field is used for version management in production tracking, quality control, and subsequent iterative design. Downstream packaging layout software can read this structured output file, automatically generating die-cutting lines and layout patterns by parsing geometric parameters, and automatically generating printing color separation files and proofing files by parsing color parameters. 3D modeling software can automatically build a 3D model based on outline control points and size parameters and assign materials and colors. The printing plate-making system can automatically generate plate-making process parameters based on color coordinates and material information. This allows for the automated implementation and mass production of complete gift set color schemes without the need for manual re-entry of design data.

[0081] It should be noted that, in this document, relational terms such as "first" and "second" are used merely 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 a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0082] 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 color-type linkage design method for unified style in gift sets, characterized in that: The color-type linkage design method includes the following steps: SP1: Obtain the design requirements and style intent of gift sets, collect the application scenarios, target audience, gift category composition and style keywords of the target gift sets, and map the style keywords into a set of style tags with weights based on the pre-built style knowledge graph. Sp2. Based on the set of style tags, construct corresponding color style vectors and shape style vectors. The color style vectors are defined in a preset color space and are used to represent the hue, brightness, saturation and color matching mode of the main color, secondary color and embellishment color of the gift. The shape style vectors are used to represent at least one of the curvature characteristics, aspect ratio, outline complexity, symmetry and texture density of the gift's appearance. Sp3. Based on the color style vector and the shape style vector, establish a color-shape linkage constraint model between the color and shape of the gift. The constraint model includes color-shape linkage rules based on style semantics and cross-category style consistency constraints for each sub-item of the gift set. Sp4. Utilize the color-type linkage design solution engine to generate candidate complete color-type design schemes; Sp5. Conduct a unified style quantitative evaluation of each candidate complete set of color design schemes to obtain a comprehensive style unity index, which includes visual unity index, color unity index and shape unity index. Based on the comprehensive style unity index, screen or iteratively optimize the candidate complete set of color design schemes. Sp6. Output at least one set of color and shape design schemes for a complete set of gifts that meet the preset uniformity threshold, and record the color parameters and shape parameters of each gift item in a parametric form.

2. The color and pattern linkage design method for unified style in gift sets according to claim 1, characterized in that: In Sp1, the construction of the style tag set includes: By expanding the nodes of style keywords in the style knowledge graph, candidate style tags associated with festival types, cultural attributes, aesthetic preferences, and target audiences can be obtained. Based on a pre-trained style tag relevance model or interactive adjustments by designers, weights are assigned to candidate style tags to form a normalized set of style tags. The style tag set is used to drive the determination of the target area for the color style vector and the styling style vector.

3. The color and pattern linkage design method for unified style in gift sets according to claim 1, characterized in that: In Sp2, the construction of the color style vector includes: In the CIELab color space, the colors in the preset color library are encoded to form a set of color vectors containing primary colors, secondary colors, and accent colors; Based on the color vector set, add brightness level relationship parameters, saturation distribution parameters, and color harmony mode parameters; Based on the style tag set and the preset color-style mapping function, the overall style of the gift set is mapped to a color style target vector region, and corresponding feasible color vector sub-regions are assigned to different gift categories in the gift set.

4. The color and pattern linkage design method for unified style in gift sets according to claim 1, characterized in that: In Sp2, the construction of the styling vector includes: For different gift categories, preset shape parameter templates are provided. The shape parameter templates include one or any combination of the following: control point coordinates of the outer contour of the gift, corner radius, length-width-height ratio, number of contour line nodes, and texture distribution density. The styling parameter template is mapped to a styling style vector, and the target value range and allowable offset range of each styling parameter are determined according to the style tag set to obtain the styling style target vector of each gift category.

5. The color and pattern linkage design method for unified style in gift sets according to claim 1, characterized in that: In Sp3, the color-type linkage constraint model includes: Color-shape semantic linkage rules are used to constrain the range of variation in curvature features, outline complexity, and texture density of the gift's appearance under a specific main color or color scheme. Cross-category style consistency constraint is used to limit the distance between the color style vector and the shape style vector of each gift item in the gift set and the style center vector of the gift set, respectively, to no more than a preset threshold, and different offset tolerances are set for the main gift and the auxiliary gift. Process and material constraints are used to limit the feasible set of color and shape parameters within the range of printing process, material strength, and dimensional specifications.

6. The color and pattern linkage design method for unified style in gift sets according to claim 1, characterized in that: In Sp4, the color-type linkage design solution engine is implemented by constructing a multi-objective optimization function, which includes: The sub-objective function representing style consistency is used to measure the degree to which the color style vector and shape style vector of each gift item in the gift set deviate from the style center vector; A sub-objective function characterizing the degree to which color-shape linkage rules are satisfied, used to measure the proportion or degree to which color-shape semantic linkage rules are satisfied; A sub-objective function characterizing aesthetic harmony is used to score a complete set of color scheme design schemes based on a pre-trained aesthetic evaluation model or color harmony formula. Sub-objective functions characterizing cost and process feasibility; The color-type linkage design solution engine uses one or any combination of genetic algorithm, simulated annealing algorithm or particle swarm optimization algorithm to solve the above multi-objective optimization function to obtain the candidate complete color-type design scheme.

7. The color and pattern linkage design method for unified style in gift sets according to claim 1, characterized in that: In Sp5, the unified style quantitative evaluation of candidate complete color scheme design schemes includes: Input the 2D renderings or 3D renderings of each gift item in the gift set into a pre-trained visual style recognition network to extract visual style feature vectors; The visual style feature vector is used to calculate the feature similarity distribution within the gift set to obtain a visual uniformity index. The visual unity index is weighted and fused with the color unity index and the form unity index calculated based on the color style vector and the form style vector to form a comprehensive style unity index.

8. The color and pattern linkage design method for unified style in gift sets according to claim 1, characterized in that: The Sp5 further includes: When the overall style consistency index is lower than the preset consistency threshold, an interactive optimization process is triggered, displaying key color or shape parameters that have a significant impact on consistency to the designer. The system receives adjustment instructions from the designer for the key parameters and, based on the adjusted parameters, calls the color-type linkage design solving engine again for local optimization to generate new candidate complete color-type design schemes.

9. A color-type linkage design method for unified style in gift sets according to any one of claims 1-8, characterized in that: The system corresponding to the color-type linkage design method includes the following components: The style intent acquisition module is used to obtain the design requirements and style intent for gift sets and generate a set of style tags; The style vector construction module is used to construct color style vectors and styling style vectors based on the style tag set; The color-shape linkage constraint construction module is used to establish a color-shape linkage constraint model for color-shape linkage rules and cross-category style consistency constraints; The color and shape linkage design solution module is used to jointly optimize the color parameters and shape parameters of each gift item in the gift set under the color and shape linkage constraint model, and generate candidate complete color and shape design schemes. The unified style evaluation module is used to quantitatively evaluate the unified style of candidate complete color scheme design schemes and output a comprehensive style unification index. The solution output module is used to output a set of color and shape design schemes for gifts that meet the preset uniformity threshold, along with their corresponding parametric color parameters and shape parameters.

10. A color-type linkage design method for unified style in gift sets according to claim 1, characterized in that: The system corresponding to the color-type linkage design method also includes a parametric design data generation module, used to convert the color parameters and shape parameters in the selected complete set of color-type design schemes into structured output files that can be called by downstream design or production software. The structured output files include: CIELab color coordinates of the primary, secondary, and accent colors of each gift item and their regional distribution data on the gift surface; Parametric description of shape parameters such as contour control point coordinates, fillet radius, aspect ratio, texture density, and shape complexity; The overall style center vector of the gift set, the unified style index, and the style version number for production tracking; The structured output file can be automatically parsed by packaging layout software, 3D modeling software, or printing plate-making system, and is used to realize the automated implementation and mass production of gift set color and pattern design.

Images