Efficient optimization processing system and method for intelligently generating embroidery pattern

Abstract

The invention relates to the technical field of embroidery automation, in particular to an efficient optimization processing system and method for intelligently generating embroidery patterns, and the method comprises the following steps: collecting different styles of embroidery patterns, and establishing a style feature library; receiving the embroidery pattern and the style keyword expected to be fused, and generating the embroidery pattern by using a content recommendation algorithm; according to the pattern complexity, the fabric attributes and the color distribution, recommending an optimal stitching method and a line trend, and optimizing a stitching method combination and an embroidery path by using a reinforcement learning technology; performing analogue simulation, evaluating the color accuracy of the embroidery pattern through a bidirectional reflection distribution function and a chromatic aberration formula, calculating the stitch density, calculating the feature similarity of the generated pattern and the target style pattern, and analyzing the style consistency; and adjusting the pattern according to a simulation result to obtain a final embroidery pattern. According to the method, the intelligent style fusion model is built, the optimal fusion proportion is calculated, the unique first draft of the fusion style embroidery pattern is generated, and a designer is assisted to quickly realize style innovation.

Description

Technical Field

[0001] The present invention relates to the technical field of embroidery automation, and particularly to an efficient optimization processing system and method for intelligently generating embroidery patterns. Background Art

[0002] In the field of traditional embroidery pattern creation, the inspiration sources of designers are mainly based on their own knowledge reserves and past experiences, lacking systematic creative inspiration tools. Facing the vast and ever-changing market demands, it is difficult to conceive novel and unique patterns in a short period of time, resulting in a serious homogenization phenomenon in design works and being unable to meet consumers' pursuit of personalized and trendy embroidery patterns. Conventional pattern generation methods are unable to cope when dealing with complex graphics and detailed expressions. For embroidery patterns with fine textures, gradient effects or complex geometric shapes, the generated patterns often have problems such as edge jaggedness, detail loss, and scale imbalance, affecting the aesthetics and artistic value of the patterns and being difficult to meet the strict standards in the production of high-end embroidery products.

[0003] Currently, for the optimization of embroidery patterns, general image processing means are mostly adopted, without fully considering the characteristics of embroidery techniques and the actual application scenarios of patterns. For example, the requirements of different stitch types for the line thickness and direction, and the influence of different fabrics on the color presentation effect are not combined, resulting in situations such as stitch mismatch and color distortion in the actual embroidery process of the optimized patterns, reducing production efficiency and increasing costs. From pattern design to actual embroidery production, it involves multiple links and different professionals. Due to the lack of an effective collaboration platform and unified data standards, information is prone to deviation and omission during the transmission process. Poor communication between designers and embroidery workers leads to inaccurate conveyance of design intentions, and frequent rework phenomena occur during the production process, seriously affecting the project progress and product quality. Summary of the Invention

[0004] The present invention can generate unique fusion-style patterns according to the input style, helping designers quickly achieve style innovation.

[0005] The technical solution proposed by the present invention is: an efficient optimization processing method for intelligently generating embroidery patterns, the method comprising: Collecting embroidery patterns of different styles, performing in-depth feature extraction and analysis, and establishing a style feature library; Receiving an embroidery pattern and style keywords to be fused, and generating an embroidery pattern using a content recommendation algorithm; According to the pattern complexity, fabric properties and color distribution, recommending the optimal stitch type and line direction, and using reinforcement learning technology to optimize the combination of stitch types and embroidery paths; Simulate the generated embroidery pattern, evaluate the color accuracy of the embroidery pattern through the bidirectional reflectance distribution function and the color difference formula, calculate the stitch density, and analyze the style consistency by calculating the feature similarity between the generated pattern and the target style pattern; Adjust the pattern color according to the color accuracy, adjust the number of stitches according to the stitch density, and adjust the generated pattern according to the style consistency to obtain the final embroidery pattern.

[0006] Preferably, the deep feature extraction process includes the following steps: Use a convolutional neural network to extract the line direction, line thickness, and number of lines to obtain a feature map; use an edge detection algorithm to extract the edge information of the image; use a gray-level co-occurrence matrix to extract texture features; count the histograms of different color components in the HSV color space; Fuse the stitch, line, fabric, and color features to generate a comprehensive feature vector , where and are the feature vectors obtained from edge detection and texture analysis respectively, is the feature vector of the fabric-color combination, is the feature map.

[0007] Preferably, the process of generating the embroidery pattern is as follows: Analyze the embroidery pattern input by the user according to the pattern complexity feature, fabric attribute feature, and color distribution feature to generate a user style preference vector; according to the analysis result, transfer the target style to the input embroidery pattern based on the generative adversarial network to generate a new embroidery pattern; calculate the optimization loss for the generated embroidery pattern.

[0008] Preferably, the new embroidery pattern is generated by balancing the content loss and the style loss; the content loss is obtained from the mean square error between the generated pattern and the content pattern in the feature space; the style loss is obtained by calculating the difference between the Gram matrices of the generated pattern and the style pattern.

[0009] Preferably, the content recommendation algorithm recommends patterns with high similarity by calculating the similarity between the user style preference vector and the comprehensive feature vector, and the similarity is calculated by the cosine similarity method. The specific formula is as follows: Let the user's style preference vector be: ; The style feature vector of the pattern is: ; Calculate the dot product of the user style preference vector and the pattern style feature vector: ; Calculate the modulus of the user style preference vector: ; Calculate the modulus of the pattern style feature vector: ; The similarity is calculated using cosine similarity: ; Where: represents the number of features, represents the preference value of the user in the th dimension, represents the feature value of the pattern in the th feature dimension.

[0010] Preferably, the bidirectional reflectance distribution function calculates the reflected light intensity through ambient light, diffuse reflection, and specular reflection; different light conditions are simulated by changing the light source intensity, direction, and color to obtain the final reflected light color.

[0011] Preferably, the color difference formula is calculated through the corrected brightness, chromaticity, and hue differences to analyze the color difference between the generated pattern color and the target color. The specific content is as follows: Let the color of the generated pattern be , and the target color be , then the color difference is: ; Where, , , are the corrected brightness, chromaticity, and hue differences respectively; , , are the corresponding weight factors respectively; is the hue rotation factor; , , are parameters with a value of 1.

[0012] Preferably, the analysis process of the style consistency is as follows: Construct a prediction model through a convolutional neural network; input the generated embroidery pattern, pass through the convolutional layer, pooling layer, and fully connected layer, and output the predicted value of the quality evaluation index; use the mean square error loss function to calculate the error between the predicted value and the true value.

[0013] Preferably, the style feature library adopts a distributed storage technology to store the embroidery pattern data on the blockchain; homomorphic encryption is used during data storage and operation.

[0014] The present invention also provides an efficient optimization processing system for intelligently generating embroidery patterns, and the system is used to execute the efficient optimization processing method for intelligently generating embroidery patterns as described above.

[0015] Advantages of the present invention: By using transfer learning technology, an intelligent style fusion model is built to pre-extract and analyze deep features of embroidery pattern data in different styles, and a style feature library is established. When the user inputs style keywords expected to be fused, the model automatically extracts corresponding style features from the feature library, calculates the optimal fusion ratio through algorithms, and generates a preliminary embroidery pattern with a unique fusion style, helping designers quickly achieve style innovation. An embroidery pattern resource management system is constructed to uniformly classify, label, and store various resources such as materials and models. By using intelligent search algorithms, designers only need to input simple keywords, and the system can quickly and accurately match the required resources and provide information such as resource usage frequency and recommended similar resources, effectively improving resource utilization rate and saving design time.

[0016] At the data storage level, blockchain technology is used to distribute the storage of pattern data, and homomorphic encryption algorithms are used to ensure the security of data during storage and operation. Even if the data is stolen, it cannot be cracked. During the data transmission process, the SSL / TLS encryption protocol is used to prevent data from being eavesdropped and tampered with, comprehensively ensuring data security. Description of the Drawings

[0017] Figure 1 It is a flowchart of an efficient optimization processing method for intelligently generating embroidery patterns according to the present invention; Figure 2 It is a flowchart of generating an embroidery pattern of an efficient optimization processing method for intelligently generating embroidery patterns according to the present invention. Detailed Embodiments

[0018] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles defined in the following description can be applied to other implementation schemes, variant schemes, improvement schemes, equivalent schemes, and other technical schemes that do not deviate from the spirit and scope of the present invention.

[0019] It can be understood that the term "one" should be understood as "at least one" or "one or more". That is, in one embodiment, the number of an element can be one, and in other embodiments, the number of the element can be multiple. The term "one" cannot be understood as a limitation on the number.

[0020] Such as Figure 1 And Figure 2As shown, collect embroidery patterns of different styles (such as traditional embroidery, modern embroidery, etc.), including pattern images, stitch descriptions, line directions, fabric types, and color matching information. Mark the feature points, line thicknesses, stitch types, fabric properties, and color distributions of the patterns.

[0021] Stitch and line feature extraction: Use CNN to extract the stitch and line features of the pattern, including line thickness, direction, density, etc. Through edge detection and texture analysis algorithms, further extract the detailed features of the pattern.

[0022] Fabric and color feature extraction: Use color space conversion (such as RGB to HSV) to analyze the color distribution of the pattern. Combine fabric properties (such as material, texture, glossiness) to extract the combined features of fabric and color.

[0023] Feature fusion: Integrate stitch, line, fabric, and color features to generate a comprehensive feature vector. Use the attention mechanism to weight different features and highlight key features. [[ID=I1]]

[0024] The specific content is as follows: Use a convolutional neural network (CNN) to extract features from the embroidery pattern. Let the input image be , where H, W, and C are the line direction, line thickness, and number of lines respectively. CNN can be represented as a combination of a series of convolutional layers, pooling layers, and activation functions. After layers of convolutional operations, the feature map is obtained.

[0025] Use an edge detection algorithm (such as Canny edge detection) to extract the edge information of the pattern. Let the edge detection function be , then the edge image .

[0026] Texture analysis can use methods such as the gray-level co-occurrence matrix (GLCM). Let be the gray-level co-occurrence matrix, where and are gray levels, is the distance, is the direction. Texture features such as contrast, correlation, energy, and homogeneity can be extracted from the GLCM.

[0027] Convert the RGB color space to the HSV color space. Let be the RGB color vector, and the conversion function be , then .

[0028] Analyze the color distribution of the pattern by statistically analyzing the histograms of different color components in the HSV color space. Let , and are histograms of hue, saturation, and brightness, respectively.

[0029] Let the fabric attribute vector be , where is the material, is the texture, is the glossiness. The fabric attributes and color features are fused to generate the combined feature vector of fabric and color .

[0030] The stitch, line, fabric, and color features are fused to generate the comprehensive feature vector , where and are feature vectors obtained from edge detection and texture analysis, respectively.

[0031] The attention mechanism is used to weight different features to highlight key features. Let the weight vector of the attention mechanism be , where n is the number of features. Then the weighted feature vector is , where is the -th feature vector. The attention weight can be calculated by the following formula: ; where is the score of the -th feature, which can be calculated by a fully connected layer: ; where and are the weights and biases of the fully connected layer, respectively.

[0032] A style transfer model is established: Based on the generative adversarial network (GAN) or neural style transfer technology, the target style (such as modern art style) is transferred to the input pattern. By optimizing the loss function (such as content loss, style loss), it is ensured that the generated pattern incorporates the target style while retaining the content of the original image.

[0033] The features of multiple embroidery styles are fused to generate a new style. Using multi-task learning technology, the style fusion and pattern generation tasks are optimized simultaneously to form a style fusion model. The specific content is as follows: Use the generative adversarial network (GAN) or neural style transfer technology to transfer the target style to the input pattern. Let the generator be G and the discriminator be D (for GAN), the content image be , and the style image be . For neural style transfer, the generated image It can be obtained by minimizing the content loss and the style loss as follows: ; where and are hyperparameters used to balance the content loss and the style loss. The content loss can be defined as the mean squared error between the generated image and the content image in the feature space: ; where and are the feature maps of the generated image and the content image at a certain layer respectively. The style loss can be obtained by calculating the difference between the Gram matrices of the generated image and the style image: ; where and are the Gram matrices of the generated image and the style image at the th layer respectively, and are the height and width of the feature map at the th layer respectively.

[0034] Use multi-task learning technology to optimize the style fusion and pattern generation tasks simultaneously. Let the loss of the style fusion task be , and the loss of the pattern generation task be , then the total loss is: ; where and are hyperparameters used to balance the losses of the two tasks.

[0035] Recommend the optimal stitch type and line direction according to the pattern complexity, fabric properties, and color distribution.

[0036] Use reinforcement learning technology to optimize the combination of stitch types and embroidery paths.

[0037] Recommend the optimal stitch type and line direction according to the pattern complexity, fabric properties, and color distribution. It can be achieved using a regression model or a classification model. Let the input feature vector be , where is the pattern complexity feature, is the fabric property feature, is the color distribution feature. The output of the model is the recommended stitch type and line direction.

[0038] Recommend suitable fabric and color combinations according to the user's input style preference.

[0039] Use collaborative filtering or content-based recommendation algorithms to generate personalized recommendation results.

[0040] Optimize the combination of stitch methods and embroidery paths using reinforcement learning techniques. Let the state space be S, the action space be A, and the reward function be R(s, a), where is the current state, is the current action. The goal of reinforcement learning is to maximize the cumulative reward: ; where is the discount factor, is the time step. Policy gradient algorithms (such as A2C, PPO, etc.) can be used to learn the optimal policy .

[0041] Generate personalized recommendation results according to the user's input style preferences using collaborative filtering or content-based recommendation algorithms. Let the user's style preference vector be , where represents the number of features, represents the user's preference value in the -th dimension. The style feature vector of the pattern is , represents the feature value of the pattern in the -th feature dimension. Calculate the dot product of the user's style preference vector and the pattern's style feature vector: . Calculate the norm of the user's style preference vector: , and calculate the norm of the pattern's style feature vector: .

[0042] For the collaborative filtering algorithm, calculate the similarity between users, find users similar to the target user, and then recommend the patterns liked by these users.

[0043] For the content-based recommendation algorithm, calculate the similarity between the user's style preference vector and the pattern's style feature vector, and recommend patterns with high similarity. The similarity is calculated using methods such as cosine similarity: ; Simulate the generated embroidery patterns to estimate the actual embroidery effect. Use image rendering technology to simulate the embroidery effect under different fabric and lighting conditions. Define quality evaluation metrics for embroidery patterns (such as stitch density, color accuracy, style consistency, etc.). Predict the quality of the embroidery pattern through a deep learning model and provide optimization suggestions.

[0044] Different fabrics have different reflection characteristics and textures, and the reflection of light by fabrics can be simulated through the bidirectional reflectance distribution function (BRDF). For common fabrics such as silk and cotton, existing BRDF models such as the Phong model can be used. The specific content of the Phong model is as follows: The reflected light intensity is composed of the ambient light , the diffuse reflection light and specularly reflected light The formula is as follows: ; Ambient light part: , where is the ambient light reflection coefficient, is the ambient light intensity.

[0045] Diffuse reflection part: , where is the diffuse reflection coefficient, is the light intensity, is the normal vector of the fabric surface, is the vector from the surface point to the light source.

[0046] Specular reflection part: , where is the specular reflection coefficient, is the reflected light vector, is the vector from the surface point to the observer, is the glossiness index.

[0047] By changing the intensity, direction, and color of the light source, different lighting conditions can be simulated. Assume the light source intensity is , the direction is , and the color is , then the color of the reflected light can be calculated by the following formula: For the color of the diffuse reflection part: ; For the color of the specular reflection part: ; Final reflected light color: , where is the ambient light color.

[0048] The stitch density can be defined as the number of stitches per unit area. Let the area of the pattern area be A and the number of stitches be N, then the stitch density is: The color accuracy can be measured by calculating the color difference between the color of the generated pattern and the target color. The commonly used color difference formula is the CIEDE2000 color difference formula. Let the color of the generated pattern be , and the target color be , then the color difference is: ; Among them, , , are the corrected brightness, chroma, and hue differences, , , are the corresponding weight factors, is the hue rotation factor, , , are parameters, usually taking the value of 1.

[0049] The style consistency can be measured by calculating the feature similarity between the generated pattern and the target style pattern. The feature vectors of the patterns can be extracted using a pre-trained convolutional neural network (such as VGG, ResNet), and then the cosine similarity between the feature vectors can be calculated. Let the feature vector of the generated pattern be , and the feature vector of the target style pattern be , then the style consistency is: ; Collect a large amount of embroidery pattern data, including pattern images, corresponding quality evaluation metrics (such as stitch density, color accuracy, style consistency, etc.), and divide them into a training set, a validation set, and a test set. Use a convolutional neural network (CNN) to build a prediction model. Assume that the input embroidery pattern image is , and after a series of convolutional layers, pooling layers, and fully connected layers, the predicted value of the quality evaluation metric is output.

[0050] Use the mean squared error loss function (MSE) to measure the difference between the predicted value and the true value. Let the true quality evaluation metric be , then the loss function is: ; where: is the number of samples.

[0051] Use the training set to train the model, update the model's parameters through the backpropagation algorithm, and minimize the loss function. Validate on the validation set and adjust the model's hyperparameters, such as the learning rate, batch size, etc.

[0052] If the predicted stitch density is too low, increase the number of stitches; if it is too high, decrease the number of stitches. The specific adjustment amount can be determined according to the prediction error of the stitch density and practical experience.

[0053] If the predicted color accuracy is poor, analyze the main sources of color difference, such as the brightness, chroma, or hue difference of the color. For brightness difference, adjust the lightness of the color; for chroma difference, the saturation of the color can be adjusted; for hue difference, adjust the hue of the color.

[0054] If the predicted style consistency is low, use neural style transfer technology to further adjust the style of the generated pattern to make it closer to the target style. Control the degree of style transfer by adjusting the weights of the style loss and the content loss.

[0055] Adopt distributed storage technology to store the embroidery pattern data on the blockchain to ensure the immutability and traceability of the data. First, standardize the embroidery pattern data to unify the image size, format, etc. For example, convert all embroidery pattern pictures to the PNG format with a unified size of 512x512 pixels.

[0056] Use homomorphic encryption during data storage and operation to ensure that even if the data is stolen, it cannot be cracked, thus protecting data privacy. Homomorphic encryption allows specific calculations to be performed on ciphertext without decryption. Adopt the Paillier homomorphic encryption algorithm, which supports additive and multiplicative homomorphic operations. Select IPFS (InterPlanetary File System) as the distributed storage system. IPFS is a distributed file system that stores and retrieves files through content addressing. Record the hash value of IPFS and the metadata of the data (such as pattern name, creation time, etc.) on the blockchain. Use Ethereum and the Web3.py library to interact with the Ethereum network. When performing operations on the embroidery pattern data, since the data is encrypted, homomorphic operations can be directly performed on the ciphertext. For example, adjust the brightness of the pattern. Use a server that supports SSL / TLS encryption to transfer the embroidery pattern data from one node to another. When the client requests data, it will automatically establish an SSL / TLS encrypted connection with the server to ensure the security of data transmission. When the embroidery pattern data is needed, use the private key for decryption, realizing the secure storage, operation, and transmission of the embroidery pattern data, and ensuring the immutability, traceability, and privacy of the data.

[0057] Embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. Embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), the above-mentioned functions defined in the methods of the present application are executed. It should be noted that the above-mentioned computer-readable medium in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium can include, but are not limited to: an electrical connection having one or more wire segments, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction execution system, apparatus, or device. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries the computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and the computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted by any suitable medium, including but not limited to: wireless, wire segments, optical cables, RF, etc., or any suitable combination of the above.

[0058] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that, in some alternative implementations, the functions noted in the blocks may occur in a different order than that noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0059] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functions and structural principles of the present invention have been demonstrated and explained in the embodiments. Without departing from the said principles, the embodiments of the present invention can have any variations or modifications.

Claims

1. An efficient optimization method for intelligently generating embroidery patterns, characterized in that, The method includes the following steps: Collect embroidery patterns of different styles, perform in-depth feature extraction and analysis, and establish a style feature library; Receive an embroidery pattern and style keywords to be fused, and generate an embroidery pattern using a content recommendation algorithm; Recommend the optimal stitch type and line direction based on pattern complexity, fabric properties, and color distribution, and use reinforcement learning technology to optimize the combination of stitches and the embroidery path; Perform simulation on the generated embroidery pattern, evaluate the color accuracy of the embroidery pattern through the bidirectional reflectance distribution function and color difference formula, calculate the stitch density, and analyze the style consistency by calculating the feature similarity between the generated pattern and the target style pattern; Adjust the pattern color according to the color accuracy, adjust the number of stitches according to the stitch density, and adjust the generated pattern according to the style consistency to obtain the final embroidery pattern.

2. The efficient optimization processing method for intelligently generating embroidery patterns according to claim 1, characterized in that, The in-depth feature extraction process includes the following steps: Use a convolutional neural network to extract line direction, line thickness, and number of lines to obtain a feature map; use an edge detection algorithm to extract the edge information of the image; use a gray-level co-occurrence matrix to extract texture features; and count the histograms of different color components in the HSV color space; Fuse the stitching method, lines, fabric, and color features to generate a comprehensive feature vector , where and are feature vectors obtained from edge detection and texture analysis respectively, is the matching feature vector of fabric and color, is the feature map.

3. An efficient optimization processing method for intelligently generating embroidery patterns according to claim 2, characterized in that, The process of generating an embroidery pattern is as follows: Analyze the embroidery pattern input by the user according to the pattern complexity feature, fabric property feature, and color distribution feature to generate a user style preference vector; based on the analysis result, transfer the target style to the input embroidery pattern using a generative adversarial network to generate a new embroidery pattern; Calculate the optimization loss for the generated embroidery pattern.

4. An efficient optimization processing method for intelligently generating embroidery patterns according to claim 3, characterized in that, The generation of the new embroidery pattern is achieved by balancing the content loss and the style loss; the content loss is obtained from the mean square error between the generated pattern and the content pattern in the feature space; the style loss is obtained by calculating the difference between the Gram matrices of the generated pattern and the style pattern.

5. An efficient optimization processing method for intelligently generating embroidery patterns according to claim 3, characterized in that The content recommendation algorithm recommends patterns with high similarity by calculating the similarity between the user style preference vector and the comprehensive feature vector. The similarity is calculated using the cosine similarity method, and the specific formula is as follows: Let the style preference vector of the user be: ； The style feature vector of the pattern be: ； Calculate the dot product of the user style preference vector and the pattern style feature vector: ； Calculate the norm of the user style preference vector: ； Calculate the norm of the pattern style feature vector: ； The similarity is calculated using cosine similarity: ; Wherein: represents the number of features, represents the user's preference value in the th dimension, represents the feature value of the pattern in the th feature dimension.

6. An efficient optimization processing method for intelligently generating embroidery patterns according to claim 5, characterized in that, The bidirectional reflectance distribution function calculates the reflected light intensity through ambient light, diffuse reflection, and specular reflection; different lighting conditions are simulated by changing the light source intensity, direction, and color to obtain the final reflected light color.

7. An efficient optimization processing method for intelligently generating embroidery patterns according to claim 6, characterized in that, The color difference formula calculates through the corrected brightness, chromaticity, and hue differences to analyze the color difference between the color of the generated pattern and the target color. The specific content is as follows: Let the color of the generated pattern be , and the target color be . Then the color difference is: ； wherein, , , are respectively the corrected luminance, chrominance and hue differences; , , are respectively the corresponding weight factors; is the hue rotation factor; , , are parameters with values of 1.

8. An efficient optimization processing method for intelligently generating embroidery patterns according to claim 7, characterized in that, The analysis process of the style consistency is as follows: Construct a prediction model through a convolutional neural network; input the generated embroidery pattern, pass through the convolutional layer, pooling layer, and fully connected layer, and output the predicted value of the quality evaluation index; use the mean square error loss function to calculate the error between the predicted value and the true value.

9. An efficient optimization processing method for intelligently generating embroidery patterns according to claim 1, characterized in that, The style feature library uses distributed storage technology to store embroidery pattern data on the blockchain; homomorphic encryption is used during data storage and operation.

10. An efficient optimization processing system for intelligently generating embroidery patterns, characterized in that, The system is used to execute an efficient optimization processing method for intelligently generating embroidery patterns according to any one of claims 1-9.

Images