Product appearance optimization method, device, terminal equipment and storage medium

Abstract

The present application is applicable to the field of data processing technology, and provides a product appearance optimization method, apparatus, terminal device and storage medium. The method screens out multiple core perceptual words through the subjective weight values and objective weight values of perceptual words, so that the core perceptual words can more accurately reflect the user's perceptual cognition of the product appearance. In addition, an initial neural network model is trained by designing a model training set constructed by combining features and core perceptual words to obtain a perceptual evaluation model. Preselected appearance design features are input into the perceptual evaluation model to obtain corresponding perceptual evaluation values. Then, the appearance design of the target product is adjusted according to the perceptual evaluation values. The perceptual evaluation values output by the perceptual evaluation model can be used to predict the user's perception of the product's appearance design, so that when the perceptual evaluation values are low, the product appearance can be improved in a timely manner, thereby improving the efficiency of product appearance optimization.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a product appearance optimization method, apparatus, terminal device, and storage medium. Background Art

[0002] With intensifying market competition and rising consumer demands for product design, product appearance has become a crucial factor in consumer purchasing decisions. This is particularly true in the home appliance industry, where the emotional value of product appearance directly influences consumers' emotional experience and purchasing decisions, particularly in the design process for consumer products like air purifiers, air conditioners, and smart speakers. Traditional product design methods rely primarily on the designer's experience and subjective judgment, fail to fully consider consumers' emotional needs, and lack a systematic optimization process. While traditional Kansei engineering methods can theoretically guide design, in practice, they often struggle to accurately quantify consumers' emotional needs due to a lack of effective optimization methods.

[0003] Kansei engineering, a design method that integrates consumer emotional needs with product design elements, has been widely used to optimize product appearance, color, material, and other aspects. However, in practice, the application of Kansei engineering often faces the following challenges: First, the selection and weighting of sensory terms still rely heavily on the subjective judgment of experts, lacking objective and effective evaluation methods; second, the design optimization process often fails to accurately capture the complex relationship between design parameters and consumer emotional needs, resulting in inefficient design optimization; third, traditional Kansei engineering methods often rely on manual adjustments, making it difficult to achieve large-scale and diverse design optimization. Therefore, how to accurately and comprehensively identify and optimize consumer emotional needs in product appearance design and improve the efficiency of product appearance optimization has become a technical problem that needs to be solved urgently. Summary of the Invention

[0004] The embodiments of the present application provide a product appearance optimization method, apparatus, terminal device and storage medium, which can solve the technical problems in the prior art that product appearance design cannot accurately and comprehensively identify and optimize consumers' emotional needs, and that the efficiency of product appearance optimization is low.

[0005] In a first aspect, an embodiment of the present application provides a method for optimizing product appearance, comprising:

[0006] Acquire multiple perceptual terms related to the design of the target product, and construct a judgment matrix corresponding to each perceptual term according to the relative importance score of each perceptual term in the multiple perceptual terms;

[0007] If the judgment matrix passes the consistency test, determining the subjective weight value of the corresponding perceptual vocabulary according to the judgment matrix;

[0008] Constructing an evaluation matrix based on the perceptual vocabulary scores of each sample product on each perceptual vocabulary;

[0009] Determining the objective weight value of each perceptual vocabulary according to the evaluation matrix;

[0010] Selecting a plurality of core perceptual words from the perceptual words according to the subjective weight value and the objective weight value of the perceptual words;

[0011] Constructing a model training set based on the multiple design combination features of the target product and the multiple core perceptual words;

[0012] Training an initial neural network model using the model training set to obtain a perceptual evaluation model, wherein the initial neural network model is composed of a genetic algorithm and a back propagation neural network;

[0013] Inputting a preselected design feature into the perceptual evaluation model to obtain a perceptual evaluation value corresponding to the preselected design feature;

[0014] The preselected appearance design features are adjusted according to the perceptual evaluation values to optimize the appearance of the target product.

[0015] Furthermore, when the judgment matrix passes the consistency test, determining the subjective weight value of the corresponding perceptual vocabulary according to the judgment matrix includes:

[0016] Determining the preliminary weight of each perceptual vocabulary according to the product of the elements in each row of the judgment matrix;

[0017] Normalizing the initial weights of the perceptual words to obtain initial weight values corresponding to the perceptual words;

[0018] According to the initial weight values and judgment matrix corresponding to each perceptual word, the maximum eigenvalue corresponding to the judgment matrix is determined;

[0019] Determining the consistency index of the judgment matrix according to the maximum eigenvalue value and the matrix order corresponding to the judgment matrix;

[0020] determining an adjustment coefficient of the judgment matrix according to the matrix order of the judgment matrix;

[0021] Determining a consistency ratio of the judgment matrix according to the consistency index and the adjustment coefficient of the judgment matrix;

[0022] When the consistency ratio is less than a preset ratio, the initial weight value is used as the subjective weight value of the corresponding perceptual vocabulary.

[0023] Furthermore, determining the objective weight value of each perceptual word according to the evaluation matrix includes:

[0024] Determining the mean score and standard deviation of each perceptual vocabulary according to the evaluation matrix;

[0025] Determining the coefficient of variation of each of the perceptual words according to the mean score and the standard deviation of the score of each of the perceptual words;

[0026] Normalizing the coefficient of variation of each of the perceptual words to obtain the objective weight of each of the perceptual words.

[0027] Furthermore, the method of selecting a plurality of core perceptual words from the perceptual words according to the subjective weight value and the objective weight value of the perceptual words includes:

[0028] Determining the coupling degree of each of the perceptual words according to the subjective weight value and the objective weight value of each of the perceptual words;

[0029] Determining the coupling coordination degree of each perceptual vocabulary according to the coupling degree of each perceptual vocabulary and a preset adjustment factor;

[0030] The inductive words whose coupling coordination degree is greater than the preset coupling coordination degree are determined as core inductive words.

[0031] Furthermore, constructing a model training set based on the multiple design combination features of the target product and the multiple core perceptual words includes:

[0032] Obtaining the perceptual values corresponding to the multiple design combination features of the target product on each core vocabulary;

[0033] Encode each design combination feature to obtain a coding feature;

[0034] A model training set is constructed based on the encoding features and the corresponding perceptual values.

[0035] Furthermore, the genetic algorithm is used to optimize the weights and thresholds of the initial neural network model.

[0036] Furthermore, the step of selecting a plurality of core sentiment words from the sentiment words according to the subjective weight value and the objective weight value of the sentiment words further includes:

[0037] Obtaining a subjective rating set and an objective rating set related to the target product and a plurality of perceptual words;

[0038] Calculating the comprehensive weight value of each perceptual vocabulary using a comprehensive weight calculation formula based on the subjective score set, the objective score set, the subjective weight value, and the objective weight value;

[0039] The perceptual words whose comprehensive weight value is greater than the preset weight threshold are determined as core perceptual words;

[0040] The comprehensive weight calculation formula is:

[0041]

[0042] Where, For the The comprehensive weight value of each perceptual word; is a preset coefficient, the value range is [0,1], and ; For the The subjective weight of each perceptual word; For the A user's subjective rating; For the An objective rating; is a subjective rating set; is an objective scoring set; is the average score of the subjective rating set; is the average score of the objective rating set; is the variance of the subjective rating set; is the variance of the objective score set; is the minimum score value in the subjective score set and the objective score set; is the maximum score value in the subjective score set and the objective score set.

[0043] In a second aspect, an embodiment of the present application provides a device for optimizing product appearance, comprising:

[0044] an acquisition module, configured to acquire a plurality of perceptual terms related to the appearance design of a target product, and construct a judgment matrix corresponding to each perceptual term according to a relative importance score of each perceptual term in the plurality of perceptual terms;

[0045] A first determining module is configured to determine, if the judgment matrix passes the consistency test, the subjective weight value of the corresponding perceptual vocabulary according to the judgment matrix;

[0046] The first building module is used to build an evaluation matrix based on the perceptual vocabulary scores of each sample product on each perceptual vocabulary;

[0047] A second determining module is used to determine the objective weight value of each perceptual word according to the evaluation matrix;

[0048] A screening module, configured to screen out a plurality of core perceptual words from the perceptual words according to the subjective weight value and the objective weight value of the perceptual words;

[0049] The second construction module constructs a model training set based on the multiple design combination features of the target product and the multiple core perceptual words;

[0050] A training module, configured to train an initial neural network model using the model training set to obtain a perceptual evaluation model, wherein the initial neural network model is composed of a genetic algorithm and a back propagation neural network;

[0051] an input module, configured to input a preselected design feature into the perceptual evaluation model to obtain a perceptual evaluation value corresponding to the preselected design feature;

[0052] An adjustment module is used to adjust the preselected appearance design features according to the perceptual evaluation value to optimize the appearance of the target product.

[0053] In a third aspect, an embodiment of the present application provides a terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of the first aspect above when executing the computer program.

[0054] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method of the first aspect above is implemented.

[0055] It can be understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here.

[0056] Compared with the prior art, the embodiments of the present application have the following beneficial effects: multiple core perceptual words are screened out through the subjective weight values and objective weight values of perceptual words, so that the core perceptual words can more accurately reflect the user's perceptual cognition of the product appearance. In addition, the initial neural network model is trained by designing a model training set constructed by combining features and core perceptual words to obtain a perceptual evaluation model, and the pre-selected appearance design features are input into the perceptual evaluation model to obtain corresponding perceptual evaluation values. Then, the appearance design of the target product is adjusted according to the perceptual evaluation values. The perceptual evaluation values output by the perceptual evaluation model can be used to predict the user's perception of the product's appearance design, so that when the perceptual evaluation value is low, the product appearance can be improved in a timely manner, thereby improving the efficiency of product appearance optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1 This is a flowchart of the implementation of the first embodiment of a product appearance optimization method provided by the embodiments of the present application;

[0059] Figure 2 This is a flowchart of a second embodiment of a method for optimizing product appearance provided in an embodiment of the present application;

[0060] Figure 3 This is a structural block diagram of a product appearance optimization device provided in an embodiment of the present application;

[0061] Figure 4 This is a structural block diagram of a product appearance optimization device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0062] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0063] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0064] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0065] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0066] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0067] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in yet other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0068] See also Figure 1 , Figure 1 The following is a flowchart illustrating a first embodiment of a method for optimizing product appearance according to an embodiment of the present application, including:

[0069] Step S10: Acquire a plurality of perceptual words related to the appearance design of the target product, and construct a judgment matrix corresponding to each perceptual word according to the relative importance score of each perceptual word in the plurality of perceptual words.

[0070] It should be noted that the execution subject of this embodiment may be a computing service device with data processing, network communication, and program execution functions, or an electronic device, product appearance optimization device, computer, tablet computer, etc. capable of implementing the above functions. The following uses the product appearance optimization device as an example to illustrate this embodiment and the following embodiments.

[0071] It is understood that the target product may be a product requiring design optimization. Sentimental terms may reflect consumers' emotional needs for a product's design, such as modern, fashionable, technological, high-end, warm, and elegant. Relative weight scores may be scores assigned by experts after pairwise comparisons of these sentimental terms, indicating the relative importance of each. A judgment matrix may be constructed using relative importance scores as elements. The relative weight scores of each sentimental term relative to the other sentimental terms serve as a row element in the judgment matrix.

[0072] Step S20: When the judgment matrix passes the consistency test, the subjective weight values of the corresponding perceptual words are determined according to the judgment matrix.

[0073] It is understandable that the consistency of the matrix usually appears in the hierarchical analysis method to measure the rationality of the judgment matrix. If the judgment matrix passes the consistency test, it is judged that the judgment matrix is reasonable.

[0074] Step S30: constructing an evaluation matrix based on the perceptual vocabulary scores of each sample product on each perceptual vocabulary.

[0075] It is understandable that the evaluation matrix contains the perceptual vocabulary scores of all product samples. Assume that n Sample products and m Each item in the matrix x ij Indicates the i Sample products in j A rating on a perceptual vocabulary.

[0076] Step S40: determining the objective weight value of each of the perceptual words according to the evaluation matrix.

[0077] It can be understood that the objective weight value may be a value used to objectively reflect the weight of the perceptual vocabulary.

[0078] Step S50: Filtering a plurality of core sentiment words from the sentiment words according to the subjective weight values and the objective weight values of the sentiment words.

[0079] It is understood that the core emotional words can be words that are selected to more accurately reflect consumer emotions. The comprehensive weight value of each emotional word is determined based on the subjective weight value and the objective weight value, and multiple core words are selected from each emotional word based on the size of the comprehensive weight value.

[0080] Step S60: constructing a model training set based on the multiple design combination features of the target product and the multiple core perceptual words.

[0081] It can be understood that the first step is to determine the perceptual value of each design feature combination on each core perceptual vocabulary, and then use the design combination features as input data and the perceptual value as the prediction target to construct a model training set.

[0082] Step S70: The initial neural network model is trained using the model training set to obtain a perceptual evaluation model, wherein the initial neural network model is composed of a genetic algorithm and a back propagation neural network.

[0083] Step S80: inputting the preselected design features into the perceptual evaluation model to obtain perceptual evaluation values corresponding to the preselected design features.

[0084] Step S90: adjusting the preselected appearance design features according to the perceptual evaluation values to optimize the appearance of the target product.

[0085] It can be understood that if the sensory evaluation value corresponding to the pre-selected appearance design feature is less than the preset sensory evaluation value, it is determined that the pre-selected appearance design feature does not meet the design requirements, and the appearance of the target product needs to be optimized; otherwise, it is determined that the pre-selected appearance design feature meets the design requirements.

[0086] The method provided in this embodiment screens out multiple core perceptual words through the subjective and objective weight values of perceptual words, so that the core perceptual words can more accurately reflect the user's perceptual cognition of the product appearance. In addition, an initial neural network model is trained by designing a model training set constructed by combining features and core perceptual words to obtain a perceptual evaluation model. Preselected appearance design features are input into the perceptual evaluation model to obtain corresponding perceptual evaluation values. The appearance design of the target product is then adjusted based on the perceptual evaluation values. The perceptual evaluation values output by the perceptual evaluation model can be used to predict the user's perception of the product's appearance design, so that when the perceptual evaluation value is low, the product appearance can be improved in a timely manner, thereby improving the efficiency of product appearance optimization.

[0087] In some optional implementations, step S20 can be implemented through the following steps: determining the preliminary weight of each perceptual word based on the element product of each row of elements in the judgment matrix; normalizing the preliminary weight of each perceptual word to obtain the initial weight value corresponding to each perceptual word; determining the maximum eigenroot value corresponding to the judgment matrix based on the initial weight value corresponding to each perceptual word and the judgment matrix; determining the consistency index of the judgment matrix based on the maximum eigenroot value and matrix order corresponding to the judgment matrix; determining the adjustment coefficient of the judgment matrix based on the matrix order of the judgment matrix; determining the consistency ratio of the judgment matrix based on the consistency index and adjustment coefficient of the judgment matrix; when the consistency ratio is less than a preset ratio, using the initial weight value as the subjective weight value of the corresponding perceptual word.

[0088] It will be appreciated that this embodiment uses the Analytic Hierarchy Process (AHP) to determine the subjective weights of perceptual terms. The AHP is an effective multi-criteria decision analysis method, particularly suitable for addressing both qualitative and quantitative factors in complex decision-making problems. In this embodiment, the AHP is used to determine the importance of each perceptual term in the target product's design. The weight of each perceptual term in the design is quantified through expert evaluation and the AHP method.

[0089] For example, assuming the target product is an air purifier, the process for determining subjective weights using the AHP approach is as follows: 1. Constructing an evaluation matrix: First, a panel of experts assesses the importance of perceptual terms. Based on the role of each perceptual term in the air purifier's design, they perform pairwise comparisons and construct a judgment matrix. During this process, experts use a scale of 1-9 and its reciprocal to indicate the relative importance of each perceptual term, based on their experience and design requirements. The definition of the scale values is shown in Table 1.

[0090] Table 1 Meaning of judgment matrix scale values

[0091]

[0092] 2. Solve the weight vector: After obtaining the judgment matrix, the next step is to solve the weight vector of the judgment matrix. The geometric mean method is usually used to solve the weight. First, find the product of each row element:

[0093]

[0094] Where: For the i The product of rows; is the first i Rank j Column elements; m Indicates the number of sensory words.

[0095] 3. Calculate the geometric mean: Next, based on the product of each row, calculate the geometric mean of each row, which will serve as the preliminary weight of the corresponding perceptual word. The specific formula is as follows:

[0096]

[0097] Where, For the i The initial weights of the perceptual words.

[0098] 4. Normalized weight: In order to obtain the relative weight of each perceptual word, the geometric mean needs to be normalized. The normalized calculation formula is as follows:

[0099]

[0100] Where, For the i The initial weight value of the perceptual vocabulary.

[0101] 5. Calculate the maximum eigenvalue of the judgment matrix: In the hierarchical analysis method, consistency test is one of the key steps to determine whether the judgment matrix is reasonable. First, the maximum eigenvalue of the judgment matrix needs to be calculated. According to the initial weights obtained, the maximum eigenvalue The calculation formula is:

[0102]

[0103] Where, It is the vector obtained by multiplying the judgment matrix and the initial weight.

[0104] 6. Consistency test: In the hierarchical analysis method, a consistency test is required to ensure the rationality of the judgment matrix. The consistency test is performed by calculating the consistency index. The calculation formula of the consistency index CI is as follows:

[0105]

[0106] Where, n is the order of the judgment matrix.

[0107] 7. Calculate the consistency ratio: Next, the consistency ratio (CR) is used for final judgment. The calculation formula of CR is:

[0108]

[0109] As shown in Table 2, according to the order of the judgment matrix n Find the value of the adjustment coefficient RI in the table. When CR ≤ 0.1, the consistency of the judgment matrix is considered good. Otherwise, the judgment matrix needs to be re-evaluated and adjusted.

[0110] Table 2

[0111]

[0112] In some optional implementations, step S40 can be implemented through the following steps: determining the mean score and the standard deviation of each perceptual vocabulary according to the evaluation matrix; determining the coefficient of variation of each perceptual vocabulary according to the mean score and the standard deviation of the score of each perceptual vocabulary; normalizing the coefficient of variation of each perceptual vocabulary to obtain the objective weight of each perceptual vocabulary.

[0113] It will be appreciated that this embodiment uses the coefficient of variation method to determine the objective weights of perceptual terms. The coefficient of variation method is an objective method for determining weights, widely used in multi-criteria decision-making and weight allocation in product design. This method assesses the importance of each factor by calculating the degree of dispersion (variability) of the data. Generally, factors with greater variability have a greater impact on the overall system. In this embodiment, the coefficient of variation method is used to determine the objective weights of each perceptual term in the target product design.

[0114] In one example, the steps to determine the objective weight of each perceptual word by the coefficient of variation method are as follows: 1. Construct an evaluation matrix: First, it is necessary to construct an evaluation matrix that contains the perceptual word scores of all product samples. Assume that d Air purifier product samples and m Each item in the matrix Indicates the r The sample in s A rating on a perceptual vocabulary.

[0115]

[0116] Where: r=1,2,3,…,d;s=1,2,3,…,m .

[0117] 2. Calculate the coefficient of variation: For each perceptual word, first calculate the mean of its score and standard deviation The coefficient of variation is calculated by the ratio of the standard deviation to the mean to measure the relative dispersion of the perceptual vocabulary. The specific calculation is as follows:

[0118]

[0119] Where: Indicates the s The average score of the perceptual words, d is the sample size, For the r The sample in s A rating on a perceptual vocabulary.

[0120]

[0121] Where: Indicates the s The standard deviation of a perceptual word reflects the degree of dispersion of the perceptual word in all samples.

[0122]

[0123] Where: For the s The coefficient of variation of sensory words.

[0124] 3. Calculate weights: To facilitate comparison and analysis, the coefficient of variation needs to be normalized to obtain the objective weight of each perceptual word. The normalized weight represents the relative importance of each perceptual word in the overall design. The calculation is as shown in formula (11):

[0125]

[0126] In some optional implementations, step S50 can be implemented through the following steps: determining the coupling degree of each of the sensory words based on the subjective weight value and the objective weight value of each of the sensory words; determining the coupling coordination degree of each of the sensory words based on the coupling degree of each of the sensory words and a preset adjustment factor; and determining the sensory words whose coupling coordination degree is greater than the preset coupling coordination degree as core sensory words.

[0127] It can be understood that the present embodiment uses the coupling coordination method to calculate the coupling coordination of each perceptual vocabulary. The coupling coordination method is an analytical method for measuring the degree of interaction and coordination between various factors in a system, and is usually used for comprehensive evaluation of multiple factors. In the present embodiment, the coupling coordination method is applied to screen the optimal perceptual vocabulary in the appearance design of the target product. Based on the subjective weight and the objective weight, the coordination between each perceptual vocabulary is evaluated, and those words that play a key role in the design can be identified. Vocabulary with a high degree of coordination indicates a higher degree of match between subjective and objective factors. These words can better reflect the emotional needs of consumers and are then used to optimize the appearance design of the target product.

[0128] In one example, the steps for calculating the coupling coordination degree of perceptual vocabulary using the coupling coordination degree method are as follows: 1. Calculating the coupling degree: Based on the concept of "coupling" in physics, the coupling degree is used to describe the degree of interdependence between multiple systems. In this embodiment, the subjective weight and the objective weight are treated as two systems, and the coupling degree between them is calculated. The calculation formula for the coupling degree is:

[0129]

[0130] Where: Represent the weight values of each system respectively. For the subjective weight and objective weight obtained in this embodiment, they are simplified into two systems, and the coupling degree between them is calculated using formula (13):

[0131]

[0132] Where: and represent subjective weight and objective weight respectively.

[0133] 2. Construct coupling coordination model: After obtaining the coupling degree After that, the coordination between the systems needs to be further evaluated. The coordination degree is used to measure the coordination of the system, that is, whether the interaction between different systems can reach the optimal state. Therefore, in this embodiment, a coupling coordination degree model is constructed to quantify the coordination between subjective weights and objective weights. D The specific calculation formula is:

[0134]

[0135] Where: C is the coupling degree, T The preset adjustment factor is a pre-set adjustment factor.

[0136] In this embodiment, since the subjective weight and the objective weight are equally important to the final design result, the preset adjustment factor is defined as:

[0137]

[0138] Where: α=β=0.5, so the preset adjustment factor can be expressed as a balanced weighting of the two weights.

[0139] In some optional implementations, step S60 may be implemented by the following steps: obtaining the corresponding sensory values of multiple design combination features of the target product on each core vocabulary; encoding each design combination feature to obtain a coding feature; and constructing a model training set based on the coding feature and the corresponding sensory value.

[0140] In some optional implementations, the genetic algorithm is used to optimize the weights and thresholds of the initial neural network model.

[0141] In one example, this embodiment combines a genetic algorithm (GA) with a backpropagation neural network (BPNN) to optimize product appearance design. Leveraging the GA's global search capabilities, multiple design parameters within the design space are optimized to find the optimal combination of design features. Simultaneously, leveraging the powerful learning capabilities of the BPNN, the trained model captures the complex mapping relationship between design combination features and perceptual values, enabling accurate perceptual demand prediction and design optimization. Ultimately, combining the GA and BPNN creates an optimization model capable of accurately predicting consumers' emotional needs.

[0142] Back propagation neural network is a commonly used supervised learning algorithm. By training the sample set, the weights and thresholds in the network are continuously adjusted to make the network output results as close to the true value as possible. In this method, BPNN is used to predict the perceptual value of the air purifier appearance design scheme, aiming to optimize the appearance design through training models to make it more in line with the emotional needs of consumers. BPNN consists of three main layers: input layer, hidden layer and output layer. The input layer receives input data (in this embodiment, it is the design combination features of the air purifier appearance design), the hidden layer processes the data through a nonlinear activation function, and the output layer generates the final prediction result (that is, the perceptual value of the design scheme). Input layer :Each input node represents a design feature, which is passed into the network for processing. Hidden layer :After the weighted summation of neurons in the hidden layer, the activation function is calculated and the value of each neuron is output. Output layer :The output layer generates the final prediction value, that is, the perceptual value of the air purifier appearance design. BPNN adjusts the weights and thresholds of each layer through the back propagation algorithm, that is, and In addition, each neuron in the hidden layer and output layer has its own threshold value, and the size of the weight and threshold value determines the accuracy of the output result.

[0143] A genetic algorithm is an optimization algorithm that simulates natural selection and heredity mechanisms and has powerful global search capabilities. In this example, a genetic algorithm is used to optimize the weights and thresholds of a BPNN to accelerate the training process and improve model accuracy. The GA helps select the optimal weight and threshold combination, thereby reducing BPNN training time and improving prediction accuracy. The GA is combined with the BPNN to form a GA-BP model, whose main steps are as follows:

[0144] 1. Initialize the population: The first step in the genetic algorithm is to initialize the population. In this example, each individual in the population represents a combination of weights and thresholds for a BPNN. Each individual's chromosome consists of the hidden and input layer weights, the hidden layer threshold, the hidden and output layer weights, and the output layer threshold. Each chromosome is encoded as a real number, representing the weights and thresholds in the BPNN network.

[0145] 2. Fitness evaluation: Fitness evaluation is the core step in genetic algorithms. The fitness function is defined as the inverse of the mean square error (MSE):

[0146]

[0147] Where: q is the training sample size; and Indicates the i The actual and predicted values of the output nodes.

[0148] 3. Selection operation: The selection operation is used to determine which individuals will enter the next generation based on their fitness values. The commonly used selection method is Roulette Wheel Selection, where the selection probability Proportional to fitness:

[0149]

[0150] 4. Crossover: The crossover operation simulates the process of genetic recombination, creating new individuals by exchanging genes from parent individuals. The goal of the crossover operation is to produce a superior offspring by combining the strengths of both parent generations. In this example, the crossover operation uses a real crossover method, swapping chromosomes from parent individuals at random positions to generate new offspring individuals.

[0151] 5. Mutation: Mutation is a local search strategy in genetic algorithms that introduces new features by randomly modifying the genes of certain individuals. Mutation helps maintain population diversity and avoids falling into local optimal solutions.

[0152] 6. Determine the number of hidden layer nodes: The performance of BPNN is closely related to the number of hidden layer nodes. Too many hidden layer nodes may lead to overfitting, while too few may lead to underfitting. Therefore, choosing an appropriate number of hidden layer nodes is very critical. The number of nodes is often calculated using the empirical formula:

[0153]

[0154] Where: j,n,k is the number of nodes in the hidden layer, input layer, and output layer; c is a constant in the range [0,10].

[0155] 7. Optimize BPNN weights and thresholds: Obtain the output of the hidden layer, multiply it by the weight and add the threshold, and then calculate the activation function using the purelin function to obtain the final output layer result. The sum of squared errors between the output layer result and the actual sample output value is used as the fitness value to guide the optimization of the genetic algorithm. The specific calculation formula is as follows:

[0156]

[0157] Where: , Output data for the hidden layer and output layer, For input data, are the weights and thresholds of each layer, tansig and purelin are the node transfer functions.

[0158] In the optional implementation of each embodiment of the present application, the above step S50 may include the following steps S501-S503. Figure 2 This is a flowchart for implementing the second embodiment of the product appearance optimization method provided in the embodiments of the present application.

[0159] Step S501: Obtain a subjective rating set and an objective rating set related to the target product and a plurality of perceptual words.

[0160] It is understandable that the subjective rating set can be a set of ratings given by different users to the degree of conformity of the product appearance described by each perceptual word. For example, the subjective rating set is ,in For the i The objective rating set can be a set of rating values on perceptual words obtained from product design databases, market research data, etc. For example, the objective rating set is ,in For the i An objective rating value.

[0161] Step S502: Calculating the comprehensive weight value of each perceptual vocabulary using a comprehensive weight calculation formula according to the subjective scoring set, the objective scoring set, the subjective weight value, and the objective weight value.

[0162] Step S503: Determine the sentiment words whose comprehensive weight values are greater than a preset weight threshold as core sentiment words.

[0163] The comprehensive weight calculation formula is:

[0164] (20)

[0165] Where, For the The comprehensive weight value of each perceptual word; is a preset coefficient, the value range is [0,1], and ; For the The subjective weight of each perceptual word; For the A user's subjective rating; For the An objective rating; is a subjective rating set; is an objective scoring set; is the average score of the subjective rating set; is the average score of the objective rating set; is the variance of the subjective rating set; is the variance of the objective score set; is the minimum score value in the subjective score set and the objective score set; is the maximum score value in the subjective score set and the objective score set.

[0166] It is understandable that determining the comprehensive weight value through the above method has the following advantages: 1. Comprehensive application of multi-dimensional data

[0167] Traditional methods often rely solely on subjective judgment or only briefly consider some objective factors when selecting core emotional vocabulary. This embodiment comprehensively incorporates both subjective and objective evaluation data. The subjective weight reflects the user's direct perception of the product's appearance, while the objective weight provides support from objective factors such as market and design. Furthermore, the formula incorporates the ratio of the total user ratings to the total objective data, as well as the ratio of the number of users to the amount of objective data, further connecting the subjective and objective data and enabling the combined weight to more comprehensively and comprehensively reflect the characteristics of the product's appearance. 2. Dynamic Balancing Mechanism: The preset coefficients in the formula not only have a limited range of values but can also be trained using a machine learning algorithm based on historical product appearance description data. This feature allows the formula to flexibly adjust the weights of subjective and objective factors, as well as other indicators, for different product types, ensuring accurate calculation of the combined weight in various scenarios. For example, for fashion products, the influence of the subjective weight can be appropriately increased; for functional products, the proportion of the objective weight can be increased. 3. Accurately Reflecting Data Characteristics: This embodiment incorporates statistics such as mean, variance, minimum, and maximum values to deeply explore the information behind the data. The mean reflects the central tendency of the data, the variance reflects the degree of dispersion of the data, and the maximum value highlights the boundary conditions of the data. Through a series of operations, these statistics work together to more accurately reflect the inherent characteristics of subjective and objective data, as well as the differences and connections between them, thereby improving the accuracy of the comprehensive weight value. 4. Improve model adaptability: Since the preset coefficients can be optimized through machine learning, this formula has strong adaptability. As new product appearance data continues to accumulate, the model can continue to learn and adjust to adapt to market changes and the evolution of user needs. This dynamic optimization mechanism enables this formula to play a good role in the appearance description of different industries and different types of products, effectively avoiding the limitations caused by fixed parameters.

[0168] In an example, the design of the appearance of an air purifier is used as an example for illustration. The specific implementation method is as follows. Determine the emotional vocabulary: First, collect emotional vocabulary related to the appearance design of the air purifier through literature research, expert interviews and market analysis. Secondly, organize the collected emotional vocabulary into cards, one card for each word, and indicate its source (such as literature, interviews or market analysis). Finally, organize design experts, consumer representatives and researchers to conduct group discussions, and use the principles of the KJ method to freely combine and classify these emotional vocabulary. In this process, participants divide the emotional vocabulary into different categories based on the similarity and relevance of the vocabulary, and give each category a descriptive label. These labels represent the core emotional characteristics expressed by the emotional vocabulary in the design.

[0169] During the categorization process, the research team conducted in-depth analysis of each category, removing redundant, overly abstract, or irrelevant terms. Ultimately, they identified eight core emotional terms that best reflect consumers' emotional needs. These terms will serve as key input for subsequent research, guiding the optimization of product appearance design. See Table 3 for details.

[0170] Table 3 Classification of sensory vocabulary

[0171]

[0172] Subjective Weight Analysis: To determine the importance of perceptual terms related to air purifier design in consumers' emotional needs, the analytic hierarchy process (AHP) was used to analyze the subjective weights of eight perceptual terms. For example, 10 design experts (eight industrial design professors and two product design engineers) were invited to conduct pairwise comparisons based on the importance of these perceptual terms to construct a judgment matrix. The judgment matrix was constructed by using expert scoring and the maximum direction of improvement algorithm for correction. Weights were calculated according to formulas (1)-(6) and consistency tests were performed. See Tables 4-5 for details. It can be seen that the CR values of all judgment matrices are less than 0.1.

[0173] Table 4 Consistency test results

[0174]

[0175] Table 5 Perceptual vocabulary weights

[0176]

[0177] Objective weight analysis: The coefficient of variation method can effectively evaluate the importance of each perceptual word among actual consumers by analyzing the degree of dispersion of consumer ratings, thereby providing an objective basis for assigning weights to perceptual words. In order to conduct objective weight analysis, this embodiment first collected samples of the top 50 air purifiers sold on the online shopping platform. These samples represent the air purifier products that are popular in the market and have a high level of consumer attention. Next, 20 consumers were invited to evaluate the appearance design of these 50 products. Each consumer rated the appearance design of each air purifier on each perceptual word based on their own intuitive feelings and preferences. The rating standard is 1-9 points, where 1 means "completely inconsistent" and 9 means "completely consistent". Finally, the average perceptual evaluation value of each perceptual word is calculated, and the weight is calculated according to formulas (7)-(11), as shown in Table 6.

[0178] Table 6 Consumer evaluation matrix and objective weight values

[0179]

[0180] Screening of perceptual words and analysis of coupling coordination: By combining the analysis of subjective weights and objective weights and using the coupling coordination method to evaluate perceptual words, the core perceptual words that best meet the emotional needs of consumers are screened out. The core of the coupling coordination method is to calculate the coupling degree (C), coordination index (T) and final coupling coordination degree (D) between subjective weights and objective weights. These indicators can quantify the coordination of each perceptual word in the appearance design. According to the size of the D value, the matching degree of each perceptual word can be intuitively evaluated, and the effectiveness of each perceptual word in the design can be clarified by the level classification of coupling coordination, from "severe disharmony" to "high-quality coordination", as shown in Table 7. According to the analysis of formulas (12)-(15), the four perceptual words of modern, simple, fashionable and technological show a high degree of coordination between subjective and objective evaluations, and are therefore finally retained as core perceptual words, as shown in Table 8. This screening result provides a solid theoretical basis for the subsequent optimization of the appearance design of air purifiers.

[0181] Table 7 Coupling coordination level classification standards

[0182]

[0183] Table 8 Calculation results of subjective and objective weight coupling coordination

[0184]

[0185] Product component coding and morphological analysis: To systematically analyze the design form of air purifiers, this embodiment uses morphological analysis to deconstruct the appearance of sample products and summarize their key components and morphological features. By combing through market samples, six key styling elements were extracted: top contour (A), bottom contour (B), front contour (C), operation panel contour (D), air inlet contour (E), and air outlet contour (F). Their morphological factors were classified and coded, as shown in Table 9. In the coding rules, styling elements are identified by AF as the major categories, and different sub-item morphological factors are distinguished by discrete data 1-5. These numbers not only systematize the description of morphological features, but also serve as the training set for the input end of the neural network.

[0186] Table 9 Air Purifier Parts Code

[0187]

[0188] Input Layer Indicators: Based on Table 10, 30 air purifier samples were extracted and their key design elements were numerically coded. The design elements of an air purifier primarily consist of the top profile (A), bottom profile (B), front profile (C), control panel profile (D), air inlet profile (E), and air outlet profile (F) modules. These six modules served as indicators for the neural network model's input layer, so the number of input layer neuron nodes was set to six. The four sets of perceptual vocabulary quantified values served as indicators for the neural network model's output layer, so the number of output layer neurons was set to four.

[0189] To ensure comparability between the input and output layer data, this method standardizes the input layer indicators using the min-max normalization method, as shown in Table 10. After normalization, the values of each input indicator are mapped to the range of 0 to 1, effectively eliminating differences between different magnitudes. This avoids network training errors caused by magnitude differences and ensures the accuracy and stability of the neural network model.

[0190] Table 10 Design element codes and sensory evaluation values

[0191]

[0192] Construction and training of the GA-BP model: This example uses a PyTorch-based framework to construct a hybrid optimization model (GA-BP) that integrates a BP neural network and a genetic algorithm. The topology of the BPNN is defined as follows: the number of neurons in the input layer is 6, representing the six main morphological features of the air purifier design; the number of neurons in the hidden layer is 10, and the ReLU activation function is used to introduce nonlinear mapping. The number of nodes is defined by an empirical formula, where The value is 4 to balance model complexity and generalization ability. The output layer contains 4 neurons, corresponding to the normalized evaluation values of four key perceptual terms in air purifier design, using a linear activation function. The network parameters are mapped to gene vectors through real number encoding. For a model containing 6×10+10×4+10+4=114 parameters, each individual is represented as a 114-dimensional vector , the encoding rules are ,in and are the weight matrices of the hidden layer and the output layer respectively, and is the bias term. Next, GA optimization is used. First, the population is initialized and randomly generated Individuals constitute the initial population, and each gene value follows a uniform distribution , ensuring the diversity of the initial parameter space. Then, fitness evaluation is performed, where the parameter m is the number of training samples 25; the roulette wheel selection strategy is continued to retain high fitness individuals while maintaining population diversity. In crossover and mutation, the probability of single-point crossover is used. Randomly select the cut point to exchange the parent gene fragment to generate the offspring individual. Perturb the genes of offspring , to avoid premature convergence. Set the maximum evolutionary generations , when the optimal individual fitness changes for 10 consecutive generations less than Optimization terminates when the maximum number of generations is reached. The MSE curve for the optimal individual during training shows a rapid decline in the initial stages (generations 1-10), indicating that the GA effectively explores regions of high fitness. In the mid-stage (generations 10-30), the MSE enters a local optimization phase, where the convergence rate slows. Finally, after 30 generations, the algorithm stabilizes, demonstrating that it reaches global optimality and meets industrial design requirements. The model was built using PyTorch 1.12.1, using a custom genetic algorithm operator. All experiments were conducted on an NVIDIA RTX 3090 GPU, with a single training session taking ≤ 20 seconds.

[0193] Design Practice and Verification: This example breaks down the air purifier's appearance into six basic elements, each of which is further broken down into multiple form factors. By combining these form factors in different ways, 2880 possible design combinations (4 × 3 × 5 × 4 × 3 × 4 = 2880) were ultimately obtained. These different design combinations were encoded and then fed into the input layer of the constructed GA-BP neural network model for prediction. The form factor combinations with the highest predicted values for the modern, simple, fashionable, and technological indicators were A1, B3, C2, D2, E2, and F4, with a predicted value of 0.904; A2, B1, C3, D1, E3, and F2, with a predicted value of 0.873; A4, B3, C5, D3, E1, and F4, with a predicted value of 0.951; and A2, B2, C2, D4, E1, and F1, with a predicted value of 0.892. Design proposals were developed based on the experimental results of the GA-BP user perception prediction model. The GA-BP model combines a genetic algorithm with a BP neural network to address the local optimum problem of traditional neural networks in multidimensional parameter spaces, thereby enhancing the model's global optimization capabilities. The genetic algorithm provides global search capabilities, while the BP neural network captures complex relationships in data through multi-level nonlinear mapping. Application of this method to the form design of air purifier products has significantly improved the accuracy of perceptual value prediction, validating its practical potential for application in product design optimization.

[0194] Model Performance and Design Verification: To verify the predictive power and practical application of the GA-BP neural network model in air purifier design, this method compared the model's predictions with actual consumer evaluations. Table 11 shows a comparison of the predicted perceptual values with the actual perceptual means, and the relative errors were calculated. The model's prediction accuracy was high, with the relative errors for each core perceptual vocabulary within 3%. These results demonstrate the effectiveness and reliability of the GA-BP neural network model constructed using this method for air purifier design.

[0195] Table 11 Prediction comparison results

[0196]

[0197] To validate the accuracy and effectiveness of the design solutions proposed by this method, 30 consumers were invited to conduct a perceptual evaluation of the best design. The questionnaire covered the performance of the four core perceptual terms: modern, simple, fashionable, and technological. The average evaluation results were 0.913, 0.826, 0.964, and 0.854, respectively. The predicted evaluation values showed minimal variation, with an error range of less than ±0.5. This result fully validates the high consistency between the predictions generated by the GA-BP neural network model and actual consumer feedback, demonstrating the reliability and practicality of this method for product design optimization. Based on this evaluation method, companies can more accurately design and develop new products, avoid unnecessary trial and error, better meet consumers' emotional needs, reduce R&D risks, and enhance market competitiveness. This research result builds a closer bridge between designers and consumers, promotes the effective translation of consumers' emotional needs, and ultimately promotes successful product launches.

[0198] This embodiment proposes a research method that combines the complex mapping relationship between user perception and product form. Taking an air purifier as an example, it explores how to optimize the appearance design through an effective perceptual vocabulary weight assignment method, morphological analysis, and a GA-BP neural network model. Through a comparative analysis of subjective and objective weight assignment methods, combined with the combination of AHP and CV coefficients, the superiority of this method in actual design is further verified. On this basis, this embodiment can accurately predict user preferences, improve the perceptual value of design solutions, and provide a theoretical basis and practical reference for the design of other products.

[0199] The main contribution of this embodiment is reflected in the method innovation and application value. First, by combining subjective expert assignment and consumer feedback data, a subjective and objective weighting method combining the AHP method and the CV coefficient method is adopted to provide a more stable and scientific basis for the weighting of perceptual vocabulary. Secondly, the GA-BP neural network model is adopted to combine the global optimization ability of the genetic algorithm with the nonlinear mapping ability of the BP neural network to achieve precise optimization of the product appearance design. Compared with traditional methods, the model of this study shows a strong advantage in the optimization ability of multi-dimensional parameter space, effectively solves the problem of local optimality, and improves the accuracy and reliability of the prediction results.

[0200] See also Figure 3 , Figure 3 This is a structural block diagram of a product appearance optimization device 400 provided in an embodiment of the present application. In this embodiment, the product appearance optimization device includes modules for executing Figure 1-Figure 2 Each step in the corresponding embodiment. Please refer to Figure 1-Figure 2 as well as Figure 1-Figure 2 For the convenience of explanation, only the parts related to this embodiment are shown. Figure 3 , the product appearance optimization device 400 includes:

[0201] An acquisition module 401 is configured to acquire a plurality of perceptual terms related to the design of a target product, and construct a judgment matrix corresponding to each perceptual term according to a relative importance score of each perceptual term in the plurality of perceptual terms;

[0202] A first determining module 402 is configured to determine, if the judgment matrix passes the consistency test, the subjective weight value of the corresponding perceptual vocabulary according to the judgment matrix;

[0203] The first constructing module 403 is used to construct an evaluation matrix based on the perceptual vocabulary scores of each sample product on each perceptual vocabulary;

[0204] A second determining module 404 is configured to determine the objective weight value of each of the perceptual words according to the evaluation matrix;

[0205] A screening module 405 is configured to screen out a plurality of core perceptual words from the perceptual words according to the subjective weight value and the objective weight value of the perceptual words;

[0206] A second construction module 406 is configured to construct a model training set based on the multiple design combination features of the target product and the multiple core perceptual words;

[0207] A training module 407 is used to train the initial neural network model using the model training set to obtain a perceptual evaluation model, wherein the initial neural network model is composed of a genetic algorithm and a back propagation neural network;

[0208] An input module 408 is configured to input a preselected design feature into the perceptual evaluation model to obtain a perceptual evaluation value corresponding to the preselected design feature;

[0209] The adjustment module 409 is configured to adjust the preselected appearance design features according to the perceptual evaluation values to optimize the appearance of the target product.

[0210] The device screens out multiple core perceptual words through the subjective weight values and objective weight values of the perceptual words, so that the core perceptual words can more accurately reflect the user's perceptual cognition of the product appearance. In addition, the initial neural network model is trained by designing a model training set constructed by combining features and core perceptual words to obtain a perceptual evaluation model. The preselected appearance design features are input into the perceptual evaluation model to obtain corresponding perceptual evaluation values. The appearance design of the target product is then adjusted according to the perceptual evaluation values. The perceptual evaluation values output by the perceptual evaluation model can be used to predict the user's perception of the product's appearance design, so that when the perceptual evaluation value is low, the product appearance can be improved in a timely manner, thereby improving the efficiency of product appearance optimization.

[0211] It should be understood that Figure 3 In the structural block diagram of the product appearance optimization device shown in FIG, each module is used to perform Figure 1-Figure 2 The steps in the corresponding embodiments, and Figure 1-Figure 2 Each step in the corresponding embodiment has been explained in detail in the above embodiment. Figure 1-Figure 2 as well as Figure 1-Figure 2 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0212] Figure 4 This is a structural block diagram of a terminal device provided by another embodiment of the present application. Figure 4 As shown, the terminal device 500 of this embodiment includes: a processor 501, a memory 502, and a computer program 503 stored in the memory 502 and executable on the processor 501, such as a program for a product appearance optimization method. When the processor 501 executes the computer program 503, the steps in each embodiment of the above-mentioned product appearance optimization method are implemented, such as Figure 1 Alternatively, the processor 501 executes the computer program 503 to implement the above Figure 3 For details on the functions of each module in the corresponding embodiment, please refer to Figure 3 The relevant descriptions in the corresponding embodiments are not repeated here.

[0213] Exemplarily, the computer program 503 may be divided into one or more units, one or more of which are stored in the memory 502 and executed by the processor 501 to complete the present application. The one or more units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 503 in the terminal device 500.

[0214] The terminal device may include, but is not limited to, a processor 501 and a memory 502. Those skilled in the art will appreciate that Figure 4 It is only an example of the terminal device 500 and does not constitute a limitation of the terminal device 500. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the turntable terminal device may also include input and output terminal devices, network access terminal devices, buses, etc.

[0215] The processor 501 may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0216] Memory 502 can be an internal storage unit of terminal device 500, such as a hard drive or memory of terminal device 500. Memory 502 can also be an external storage terminal device of terminal device 500, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. equipped with terminal device 500. Furthermore, memory 502 can include both an internal storage unit of terminal device 500 and an external storage terminal device. Memory 502 is used to store computer programs and other programs and data required by the turntable terminal device. Memory 502 can also be used to temporarily store data that has been output or is about to be output.

[0217] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0218] If the integrated module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be either non-volatile or volatile. Based on this understanding, the present application can implement all or part of the process steps in the above-mentioned method embodiments by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. Computer-readable storage media can include: any entity or device capable of carrying computer program code, recording medium, USB flash drive, mobile hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium.

[0219] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for optimizing product appearance, characterized in that: The method comprises: Acquire multiple perceptual terms related to the design of the target product, and construct a judgment matrix corresponding to each perceptual term according to the relative importance score of each perceptual term in the multiple perceptual terms; If the judgment matrix passes the consistency test, determining the subjective weight value of the corresponding perceptual vocabulary according to the judgment matrix; Constructing an evaluation matrix based on the perceptual vocabulary scores of each sample product on each perceptual vocabulary; Determining the objective weight value of each perceptual vocabulary according to the evaluation matrix; Selecting a plurality of core perceptual words from the perceptual words according to the subjective weight value and the objective weight value of the perceptual words; Constructing a model training set based on the multiple design combination features of the target product and the multiple core perceptual words; Training an initial neural network model using the model training set to obtain a perceptual evaluation model, wherein the initial neural network model is composed of a genetic algorithm and a back propagation neural network; Inputting a preselected design feature into the perceptual evaluation model to obtain a perceptual evaluation value corresponding to the preselected design feature; Adjusting the preselected design features according to the perceptual evaluation values to optimize the appearance of the target product; The method of selecting a plurality of core perceptual words from the perceptual words according to the subjective weight value and the objective weight value of the perceptual words comprises: Determining the coupling degree of each of the perceptual words according to the subjective weight value and the objective weight value of each of the perceptual words; Determining the coupling coordination degree of each perceptual vocabulary according to the coupling degree of each perceptual vocabulary and a preset adjustment factor; Determine the perceptual words whose coupling coordination degree is greater than the preset coupling coordination degree as core perceptual words; or Obtaining a subjective rating set and an objective rating set related to the target product and a plurality of perceptual words; Calculating the comprehensive weight value of each perceptual vocabulary using a comprehensive weight calculation formula based on the subjective score set, the objective score set, the subjective weight value, and the objective weight value; The perceptual words whose comprehensive weight value is greater than the preset weight threshold are determined as core perceptual words; The comprehensive weight calculation formula is: Where, For the The comprehensive weight value of each perceptual word; is a preset coefficient, the value range is [0,1], and ; For the The subjective weight of each perceptual word; For the The objective weight of each perceptual word; For the A user's subjective rating; For the An objective rating; is a subjective rating set; is an objective scoring set; is the average score of the subjective rating set; is the average score of the objective rating set; is the variance of the subjective rating set; is the variance of the objective score set; is the minimum score value in the subjective score set and the objective score set; is the maximum score value in the subjective score set and the objective score set.

2. The method according to claim 1, wherein When the judgment matrix passes the consistency test, determining the subjective weight value of the corresponding perceptual vocabulary according to the judgment matrix includes: Determining the preliminary weight of each perceptual vocabulary according to the product of the elements in each row of the judgment matrix; Normalizing the initial weights of the perceptual words to obtain initial weight values corresponding to the perceptual words; According to the initial weight values and judgment matrix corresponding to each perceptual word, the maximum eigenvalue corresponding to the judgment matrix is determined; Determining the consistency index of the judgment matrix according to the maximum eigenvalue value and the matrix order corresponding to the judgment matrix; determining an adjustment coefficient of the judgment matrix according to the matrix order of the judgment matrix; Determining a consistency ratio of the judgment matrix according to the consistency index and the adjustment coefficient of the judgment matrix; When the consistency ratio is less than a preset ratio, the initial weight value is used as the subjective weight value of the corresponding perceptual vocabulary.

3. The method according to claim 1, wherein Determining the objective weight value of each perceptual word according to the evaluation matrix includes: Determining the mean score and standard deviation of each perceptual vocabulary according to the evaluation matrix; Determining the coefficient of variation of each of the perceptual words according to the mean score and the standard deviation of the score of each of the perceptual words; Normalizing the coefficient of variation of each of the perceptual words to obtain the objective weight of each of the perceptual words.

4. The method according to claim 1, wherein The step of constructing a model training set based on the multiple design combination features of the target product and the multiple core perceptual vocabulary comprises: Obtaining the perceptual values corresponding to the multiple design combination features of the target product on each core vocabulary; Encode each design combination feature to obtain a coding feature; A model training set is constructed based on the encoding features and the corresponding perceptual values.

5. The method according to any one of claims 1 to 4, characterized in that The genetic algorithm is used to optimize the weights and thresholds of the initial neural network model.

6. A product appearance optimization device, characterized in that: The device comprises: an acquisition module, configured to acquire a plurality of perceptual terms related to the appearance design of a target product, and construct a judgment matrix corresponding to each perceptual term according to a relative importance score of each perceptual term in the plurality of perceptual terms; A first determining module is configured to determine, if the judgment matrix passes the consistency test, the subjective weight value of the corresponding perceptual vocabulary according to the judgment matrix; The first building module is used to build an evaluation matrix based on the perceptual vocabulary scores of each sample product on each perceptual vocabulary; A second determining module is used to determine the objective weight value of each perceptual word according to the evaluation matrix; A screening module, configured to screen out a plurality of core perceptual words from the perceptual words according to the subjective weight value and the objective weight value of the perceptual words; A second construction module is configured to construct a model training set based on the multiple design combination features of the target product and the multiple core perceptual words; A training module, configured to train an initial neural network model using the model training set to obtain a perceptual evaluation model, wherein the initial neural network model is composed of a genetic algorithm and a back propagation neural network; an input module, configured to input a preselected design feature into the perceptual evaluation model to obtain a perceptual evaluation value corresponding to the preselected design feature; an adjustment module, configured to adjust the preselected design features according to the perceptual evaluation values to optimize the appearance of the target product; The screening module is further configured to determine the coupling degree of each of the perceptual words according to the subjective weight value and the objective weight value of each of the perceptual words; Determining the coupling coordination degree of each perceptual vocabulary according to the coupling degree of each perceptual vocabulary and a preset adjustment factor; Determine the perceptual words whose coupling coordination degree is greater than the preset coupling coordination degree as core perceptual words; or Obtaining a subjective rating set and an objective rating set related to the target product and a plurality of perceptual words; Calculating the comprehensive weight value of each perceptual vocabulary using a comprehensive weight calculation formula based on the subjective score set, the objective score set, the subjective weight value, and the objective weight value; The perceptual words whose comprehensive weight value is greater than the preset weight threshold are determined as core perceptual words; The comprehensive weight calculation formula is: Where, For the The comprehensive weight value of each perceptual word; is a preset coefficient, the value range is [0,1], and ; For the The subjective weight of each perceptual word; For the The objective weight of each perceptual word; For the A user's subjective rating; For the An objective rating; is a subjective rating set; is an objective scoring set; is the average score of the subjective rating set; is the average score of the objective rating set; is the variance of the subjective rating set; is the variance of the objective score set; is the minimum score value in the subjective score set and the objective score set; is the maximum score value in the subjective score set and the objective score set.

7. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images