Training and application method and system of complexity evaluation system, equipment and medium

Abstract

The invention relates to the field of mechanical design and manufacturing, and provides a construction and application method and system of a complexity evaluation system, equipment and a medium. According to the invention, an initial complexity evaluation system is constructed based on a random forest method, and the obtained geometric data of the geometric model is utilized to construct a data set to carry out multiple times of iterative training on the initial complexity evaluation system until convergence; the complexity evaluation system can evaluate the complexity of the geometric model more efficiently, objectively, accurately and universally, so that the production efficiency and quality of process products are improved.

Description

Technical Field

[0001] The present application relates to the field of mechanical design and manufacturing, and in particular to the training and application methods, systems, equipment, and media of complexity assessment systems. Background Art

[0002] In the field of mechanical design and manufacturing, it is crucial to perform performance simulation tests on the geometric models corresponding to process products. The complexity of the geometric model directly affects the computing resource requirements, processing efficiency, and final application performance of the CAE software during simulation, thereby affecting the production efficiency and quality of the process products. Geometric models that are too complex may lead to extended rendering times, increased memory usage, and delays in real-time applications, while oversimplification may affect the accuracy and visual effects of the geometric model. Therefore, by evaluating the complexity of the geometric model, the relationship between model accuracy and computational efficiency can be effectively balanced, ensuring that resource utilization is optimized and the user experience is improved while meeting application requirements.

[0003] The inventors have found that there are currently several methods for evaluating the complexity of geometric models: (1) Experience-based manual evaluation methods: Engineers make subjective judgments on the complexity of geometric models based on their own experience. This method is highly dependent on personal experience and professional knowledge and lacks objectivity and consistency. Different engineers may have different evaluations of the complexity of the same geometric model. Moreover, for complex and large-scale geometric models, manual evaluation is inefficient and difficult to accurately quantify the complexity, which cannot meet the needs of automation and large-scale engineering analysis. (2) Evaluation methods based on simple geometric feature statistics: For example, only a few intuitive geometric features such as the number of facets, the number of vertices, or the number of entities of the geometric model are calculated to measure the complexity. However, these simple geometric features cannot fully reflect the true complexity of the geometric model. A geometric model may have a small number of facets but a complex topological structure or curvature changes. Relying solely on the number of facets will underestimate its complexity, resulting in insufficient estimation of computing resources and time in subsequent CAE analysis, affecting the accuracy and reliability of the analysis results. (3) Based on traditional machine learning methods: Some studies have attempted to use other machine learning algorithms such as single decision trees and neural networks to evaluate the complexity of geometric models. However, single decision trees are prone to overfitting and have limited generalization capabilities for complex geometric models, making them difficult to accurately handle diverse geometric models. While neural networks possess powerful learning capabilities, they require extensive training data and complex parameter tuning. Furthermore, neural mesh models lack interpretability, making it difficult to intuitively understand how neural network models determine complexity based on geometric features, hindering their application and optimization in engineering practice.

[0004] In summary, existing geometric model complexity assessment methods are unable to effectively assess the quality of geometric models in terms of accuracy, objectivity, efficiency, and versatility, resulting in reduced production efficiency and quality of process products. Therefore, a more efficient, objective, accurate, and universal geometric model quality assessment method is urgently needed to improve the production efficiency and quality of process products. Summary of the Invention

[0005] The present application provides a method, system, device, and medium for constructing and applying a complexity assessment system to solve the problem that existing geometric model complexity assessment methods are unable to perform high-performance assessments of the quality of geometric models in terms of accuracy, objectivity, efficiency, and versatility.

[0006] A first aspect of the present application provides a training method for a complexity evaluation system, wherein the complexity evaluation system is used to evaluate the complexity of a geometric model, and the method comprises:

[0007] Acquiring geometric data of a geometric model, preprocessing the geometric data, and constructing a data set based on the preprocessed geometric data;

[0008] Construct an initial complexity assessment system based on the random forest method;

[0009] The initial complexity evaluation system is trained multiple times using the data set until convergence to obtain a complexity evaluation system.

[0010] In some embodiments of the present application, the initial complexity evaluation system includes a feature extraction module, a random forest module, and a complexity evaluation module, and the steps of constructing the initial complexity evaluation system based on the random forest method include:

[0011] The feature extraction module is configured to perform a first feature extraction operation on the geometric data to obtain first feature data of the geometric model;

[0012] The random forest module is configured to perform a second feature extraction operation on the first feature data to obtain second feature data of the geometric model, and calculate a complexity category prediction value of the geometric model based on the second feature data;

[0013] The complexity evaluation module is configured to calculate a complexity evaluation result of the geometric model based on the complexity category prediction value of the geometric model.

[0014] In some embodiments of the present application, the initial complexity evaluation system further includes a standardization module, and the steps of constructing the initial complexity evaluation system based on the random forest method include:

[0015] The normalization module is configured to perform normalization on the geometric data in the data set and transmit the normalization result to the feature extraction module for processing.

[0016] In some embodiments of the present application, the geometric model includes faces, edges, and vertices, and the first feature data of the geometric model includes: basic topological feature data, curvature feature data, and geometric quantity feature data, and the feature extraction module is configured to obtain the basic topological feature data, curvature feature data, and geometric quantity feature data of the geometric model in the following manner:

[0017] Count the number of faces, edges, and vertices in the geometric model, and use these data as the basic topological feature data of the geometric model;

[0018] Calculate the average curvature and Gaussian curvature of the surface in the geometric model, and count the data and area ratio of the surface in the area where the curvature of the geometric model changes dramatically, and use the average curvature and Gaussian curvature of the surface, and the data and area ratio of the surface in the area where the curvature of the geometric model changes dramatically as the curvature feature data of the geometric model;

[0019] The number and geometric parameters of holes, protrusions and grooves in the geometric model are counted, and these data are used as geometric quantity feature data of the geometric model.

[0020] In some embodiments of the present application, the random forest module includes: multiple decision tree classifiers, and the steps of constructing an initial complexity assessment system based on the random forest method include:

[0021] The plurality of decision tree classifiers are respectively configured to perform a second feature extraction operation on the first feature data to obtain a plurality of second feature data of the geometric model, and calculate a plurality of complexity category prediction values of the geometric model based on the plurality of second feature data;

[0022] The complexity evaluation module is configured to calculate multiple initial complexity evaluation results of the geometric model based on a preset category threshold and multiple complexity category prediction values, and process the multiple initial complexity evaluation results according to preset rules to obtain a complexity evaluation result of the geometric model.

[0023] In some embodiments of the present application, the data set includes: a label corresponding to the geometric model, wherein the label indicates the true value of the complexity category of the geometric model, and the steps of using the data set to iteratively train an initial complexity evaluation system until convergence to obtain the complexity evaluation system include:

[0024] During each iterative training process, the parameters of the complexity evaluation system are updated according to the following loss function:

[0025]

[0026] Where N is the number of geometric models in the dataset, y i represents the true value of the complexity category of the i-th geometric model in the dataset, Represents the complexity category prediction value of the i-th geometric model in the dataset.

[0027] A second aspect of the present application provides a complexity evaluation system trained using the complexity evaluation system training method of any of the above embodiments, the system comprising:

[0028] a feature extraction module, configured to perform a first feature extraction operation on the acquired geometric data of the geometric model to obtain first feature data of the geometric model;

[0029] a random forest module, configured to perform a second feature extraction operation on the first feature data to obtain second feature data of the geometric model, and calculate a complexity category prediction value of the geometric model based on the second feature data;

[0030] The complexity evaluation module is used to calculate the complexity evaluation result of the geometric model based on the complexity category prediction value of the geometric model.

[0031] A third aspect of the present application provides an application method of a complexity assessment system, the method comprising:

[0032] Obtaining geometric data of the geometric model;

[0033] The complexity evaluation system of the above embodiment is used to process the geometric data to obtain a complexity evaluation result of the geometric model.

[0034] The fourth aspect of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method described in any one of the first and third aspects of the above embodiments are implemented.

[0035] The fifth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method described in any one of the first and third aspects of the above embodiments is implemented.

[0036] This application has the following beneficial effects:

[0037] The above embodiment of the present application constructs an initial complexity evaluation system based on the random forest method, and uses the geometric data of the acquired geometric model to construct a data set to perform multiple iterative training on the initial complexity evaluation system until convergence, so as to obtain a complexity evaluation system that can evaluate the complexity of the geometric model more efficiently, objectively, accurately and universally, thereby improving the production efficiency and quality of process products. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and, together with the specification, are used to illustrate the technical solutions of the present application.

[0039] Figure 1 This is a flow chart of an embodiment of a training method for a complexity assessment system provided by the present application;

[0040] Figure 2 This is a schematic diagram of the framework of an embodiment of the complexity assessment system provided by this application;

[0041] Figure 3 It is a schematic diagram of the framework of another embodiment of the complexity assessment system provided by the present application;

[0042] Figure 4 This is a flow chart of an embodiment of a method for calculating characteristic data of a geometric model provided by the present application;

[0043] Figure 5 This is a flow chart of an embodiment of an application method of the complexity assessment system provided by the present application;

[0044] Figure 6 This is a schematic diagram of the framework of an embodiment of an electronic device provided by the present application;

[0045] Figure 7 It is a schematic diagram of a framework of an embodiment of a computer-readable storage medium provided by this application. DETAILED DESCRIPTION

[0046] The following describes the embodiments of the present application in detail with reference to the accompanying drawings.

[0047] In the following description, for the purpose of explanation rather than limitation, specific details such as specific system structures, interfaces, and technologies are provided to facilitate a thorough understanding of the present application.

[0048] The term "and / or" in this article is simply a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects are in an "or" relationship. In addition, "many" in this article means two or more than two. In addition, the term "at least one" in this article means any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0049] As described in the background art, existing geometric model complexity evaluation methods have the problem of being unable to perform high-performance evaluation of the quality of geometric models in terms of accuracy, objectivity, efficiency, and versatility.

[0050] In order to solve the above problems, this application proposes a new geometric model complexity evaluation scheme. In the scheme of this application, an initial complexity evaluation system is constructed based on the random forest method, and the geometric data of the acquired geometric model is used to construct a data set to perform multiple iterative training on the initial complexity evaluation system until convergence, so as to obtain a complexity evaluation system that can evaluate the complexity of the geometric model more efficiently, objectively, accurately and universally.

[0051] Before introducing the specific embodiments of the present application, we first briefly introduce some basic knowledge involved in the present application.

[0052] The geometric model is a CAD model of the process product to be produced that is exported from the CAD software.

[0053] The whole process of producing a mechanical part:

[0054] Design a part model containing holes, slots and chamfers in CAD software;

[0055] CAE software performed stress analysis on the CAD designed part model and found that the edge stress of the hole was too high;

[0056] Plan the processing technology (such as drilling, milling, chamfering) of the part model in CAPP software based on the analysis results of the part model by CAE software;

[0057] CAM software generates NC codes based on process planning and controls the machine tool to complete the processing of the part model.

[0058] Through MFR, seamless data transfer between CAD, CAE, CAPP, and CAM systems is achieved, reducing manual intervention and improving product production efficiency and quality.

[0059] The present application is described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] According to one embodiment of the present application, the present application proposes a training method for a complexity evaluation system, the complexity evaluation system is used to evaluate the complexity of a geometric model, such as Figure 1As shown, the method includes: S1, obtaining geometric data of a geometric model, preprocessing the geometric data, and constructing a data set based on the preprocessed geometric data; S2, constructing an initial complexity assessment system based on a random forest method; S3, using the data set to iteratively train the initial complexity assessment system multiple times until convergence, thereby obtaining a complexity assessment system. The geometric model of the present application can be a two-dimensional image of a CAD model exported from CAD software.

[0061] Each step is described below.

[0062] 1. Step S1

[0063] In step S1 , geometric data of a geometric model is acquired, the geometric data is preprocessed, and a data set is constructed based on the preprocessed geometric data.

[0064] According to one embodiment of the present application, geometric models include simple geometric shapes (such as spheres and cubes) and complex engineering part models (such as automotive engine components and aerospace structures). It should be noted that the structure of the geometric model is not limited to that described in the above embodiment, and the types of geometric models can be increased based on the training and application requirements of the complexity assessment system, and are not specifically limited here.

[0065] As can be seen from the above description, the above-mentioned embodiments of the present application utilize different types of geometric models to train the complexity assessment system, which can significantly improve the complexity assessment system's generalization capabilities (i.e., enhance its adaptability to different complexity levels), assessment accuracy, and breadth of application. Furthermore, the flexible adjustment and addition of geometric model types enables the complexity assessment system to continuously meet ever-changing assessment needs, thereby promoting technological innovation and cross-industry development.

[0066] 2. Step S2

[0067] In step S2, an initial complexity evaluation system is constructed based on the random forest method.

[0068] According to one embodiment of the present application, Figure 2 As shown, the initial complexity evaluation system includes a feature extraction module, a random forest module and a complexity evaluation module, and the steps of constructing the initial complexity evaluation system based on the random forest method include: the feature extraction module is configured to perform a first feature extraction operation on the geometric data to obtain first feature data of the geometric model; the random forest module is configured to perform a second feature extraction operation on the first feature data to obtain second feature data of the geometric model, and based on the second feature data, calculate the complexity category prediction value of the geometric model; the complexity evaluation module is configured to calculate the complexity evaluation result of the geometric model based on the complexity category prediction value of the geometric model.

[0069] From the above description, it can be seen that the initial complexity assessment system of the above embodiment of the present application realizes efficient analysis and complexity assessment of geometric data by integrating feature extraction module, random forest module and complexity assessment module. This method simplifies the evaluation operation process, so that even complex geometric models can be evaluated quickly and accurately, greatly promoting work efficiency and technological development in related fields.

[0070] According to one embodiment of the present application, Figure 3 As shown, the initial complexity evaluation system further includes: a standardization module, and the steps of constructing the initial complexity evaluation system based on the random forest method include:

[0071] The normalization module is configured to normalize the geometric data in the dataset and transmit the normalized results to the feature extraction module for processing. According to one embodiment of the present application, the normalization process is to unify the coordinate system and unit of the geometric data.

[0072] As can be seen from the above description, the above embodiments of the present application perform standardized processing on the geometric data, for example, unifying the coordinate system and units of the geometric data, so as to eliminate data deviations and errors caused by different coordinate systems or inconsistent units. When all geometric data are processed under a unified standard, the subsequent feature extraction process can more accurately identify and quantify the key characteristics of the geometric model, avoid additional errors that may be introduced during the conversion process, improve the data quality based on which the feature extraction module and the random forest module perform feature extraction operations, and enable the complexity evaluation module to calculate the complexity category prediction value and the final complexity evaluation result based on more accurate and representative data. Therefore, by standardizing the coordinate system and units, the entire evaluation process becomes more reliable and efficient, and the authenticity and accuracy of the complexity evaluation results of the geometric model are improved.

[0073] The feature extraction module, random forest module, and complexity evaluation module are introduced below.

[0074] (1) Feature extraction module

[0075] According to one embodiment of the present application, the geometric model includes faces, edges, and vertices, and the first feature data of the geometric model includes: basic topological feature data, curvature feature data, and geometric quantity feature data, and Figure 4As shown, the feature extraction module is configured to perform the following steps S21-S23 to obtain the basic topological feature data, curvature feature data, and geometric quantity feature data of the geometric model in the data set: S21, counting the number of faces, edges, and vertices in the geometric model, and using these data as the basic topological feature data of the geometric model; S22, calculating the average curvature and Gaussian curvature of the faces in the geometric model, and counting the data and area proportion of the faces in the area where the curvature of the geometric model changes dramatically, and using the average curvature and Gaussian curvature of the faces, and the data and area proportion of the faces in the area where the curvature of the geometric model changes dramatically as the curvature feature data of the geometric model; S23, counting the number and geometric parameters of holes, protrusions, and grooves in the geometric model, and using these data as the geometric quantity feature data of the geometric model. Among them, the curvature feature data can measure the degree of curvature and the complexity of changes of the surface of the geometric model.

[0076] According to one embodiment of the present application, the first feature data of the geometric model is a feature vector composed of basic topological feature data, curvature feature data, and geometric quantity feature data.

[0077] As can be seen from the above description, the above embodiments of the present application systematically extract the basic topological features, curvature features, and geometric quantitative features of the geometric model. That is, by counting the number of faces, edges, and vertices of the geometric model, calculating the average curvature and Gaussian curvature of each face, and counting the number and geometric parameters of holes, protrusions, and grooves, the multi-level features of the geometric model are fully captured. This method not only enhances the adaptability of the complexity assessment system to different complexity levels, but also significantly improves the accuracy and reliability of the complexity assessment system's complexity assessment of the geometric model.

[0078] According to one embodiment of the present application, S22 includes: calculating the average curvature and Gaussian curvature of the surface in the geometric model using a curvature calculation method. According to another embodiment of the present application, the curvature calculation method is a curvature calculation method based on local surface fitting.

[0079] According to one embodiment of the present application, the geometric parameters of the holes include the diameter distribution of the holes, the height range of the protrusions, etc.

[0080] From the above description, it can be seen that the above embodiment of the present application uses a curvature calculation method based on local surface fitting to calculate the average curvature and Gaussian curvature of the surface in the geometric model, so that the complexity evaluation system can more accurately capture the subtle shape features of the surface (local surface fitting can not only better handle complex and irregular surfaces, but also improve the understanding and analysis capabilities of different geometric structure features), thereby providing a more detailed geometric model description to obtain a more accurate geometric model complexity evaluation result.

[0081] (2) Random Forest Module

[0082] According to one embodiment of the present application, the random forest module includes: multiple decision tree classifiers, and the steps of constructing an initial complexity evaluation system based on the random forest method include: multiple decision tree classifiers are respectively configured to perform a second feature extraction operation on the first feature data to obtain multiple second feature data of the geometric model, and based on the multiple second feature data, multiple complexity category prediction values of the geometric model are calculated.

[0083] According to one embodiment of the present application, each decision tree classifier includes multiple nodes, and the number of features considered when each node is split is the square root of the total number of features.

[0084] As can be seen from the above description, the number of features considered by each decision tree classifier in the above embodiment of the present application when splitting a node is the square root of the total number of features. This method can achieve a good balance between ensuring the complexity of the geometric model and computational efficiency, that is, by limiting the number of features considered at each split, it not only reduces the computational burden and improves the speed of building the decision tree classifier, but also effectively avoids the overfitting problem during the training of the complexity evaluation system and enhances the generalization ability of the complexity evaluation system. In addition, this strategy helps to increase the diversity between the various decision tree classifiers. This is because different trees may be optimized based on different feature subsets when splitting nodes, thereby improving the overall accuracy and stability of the complexity evaluation system, and ultimately making the complexity evaluation results of the geometric model more reliable and accurate.

[0085] (3) Complexity Assessment Module

[0086] According to one embodiment of the present application, the complexity evaluation module is configured to calculate multiple initial complexity evaluation results of the geometric model based on a preset category threshold and multiple complexity category prediction values, and process the multiple initial complexity evaluation results according to preset rules to obtain the complexity evaluation result of the geometric model.

[0087] According to one embodiment of the present application, a preset rule is to statistically classify multiple initial complexity assessment results into complexity categories, and use the category with the largest number of categories as the complexity assessment result of the geometric model. If there are two categories with the same number of categories, a manual method is used to obtain the complexity assessment result of the geometric model.

[0088] For example, if the six initial complexity evaluation results are intermediate, intermediate, intermediate, intermediate, intermediate, and advanced, respectively, then the complexity evaluation result of the geometric model is intermediate; if the complexity evaluation results include three complexity evaluation methods, namely low, intermediate, and advanced, and if there are six decision tree classifiers in the complexity evaluation system, then the geometric model will obtain six initial complexity evaluation results. If these six initial complexity evaluation results are intermediate, intermediate, intermediate, advanced, advanced, and advanced, respectively, then the complexity of the geometric model needs to be finally classified in combination with manual methods.

[0089] As can be seen from the above description, the above embodiment of the present application utilizes multiple decision tree classifiers to perform multi-level second feature extraction operations on the first feature data, thereby generating multiple complexity category prediction values, and calculating multiple initial complexity assessment results based on these prediction values. The multiple initial complexity assessment results are then processed according to preset rules to obtain the complexity assessment result of the geometric model. This method not only improves the accuracy and robustness of complexity assessment, but also reduces the bias and misjudgment that may be caused by a single classifier by integrating the opinions of multiple decision tree classifiers, ensuring the high reliability and accuracy of the complexity assessment results.

[0090] Step S3

[0091] In step S3, the data set includes: a label corresponding to the geometric model, wherein the label indicates the true value of the complexity category of the geometric model, and the initial complexity evaluation system is iteratively trained multiple times until convergence using the data set to obtain the complexity evaluation system. The steps include:

[0092] During each iterative training process, the parameters of the complexity evaluation system are updated according to the following loss function:

[0093]

[0094] Where N is the number of geometric models in the dataset, y i represents the true value of the complexity category of the i-th geometric model in the dataset, Represents the complexity category prediction value of the i-th geometric model in the dataset.

[0095] As can be seen from the above description, the above embodiments of the present application optimize the parameters of the complexity assessment system by minimizing the loss value. This method ensures that the complexity assessment system can continuously adjust its parameters during training to reduce prediction errors, thereby improving the accuracy of the complexity assessment system's assessment of geometric model complexity. It also enhances the complexity assessment system's generalization ability, enabling it to obtain accurate quality assessment results even when processing unseen geometric models.

[0096] According to one embodiment of the present application, the complexity categories include: low, medium, and high, and the complexity category labels are 1, 2, and 3 respectively, and the complexity of each geometric model is labeled by combining expert experience evaluation and objective indicators to obtain the complexity category label of each geometric model. Experienced engineers and domain experts are organized to subjectively score the complexity of each geometric model. At the same time, the expert labeling results are calibrated and supplemented using complexity measurement indicators based on information theory (such as the information entropy of the model) to ensure the accuracy and objectivity of the labeling, thereby providing reliable training labels for the machine learning model. It should be noted that dividing the complexity categories into the above three levels is only one of the embodiments of the present application. The specific categories can be differentiated in detail according to the task requirements, and are not specifically limited here.

[0097] As can be seen from the above description, the above embodiment of the present application classifies the complexity of the geometric model and labels it as low-level (label 1), intermediate (label 2), and high-level (label 3) by combining expert experience evaluation and objective indicators based on information theory (such as the information entropy of the model), thereby ensuring the accuracy and objectivity of the complexity category labels. This method not only utilizes the profound experience and subjective judgment of experts in the field, but also calibrates and supplements the evaluation results through objective measurement indicators, thereby improving the reliability and consistency of the complexity classification. This comprehensive evaluation method provides high-quality training data for the machine learning model, enhancing the learning effect and generalization ability of the model. In addition, the ability to flexibly adjust the complexity category enables this method to adapt to different task requirements, further enhancing its scope of application and practicality.

[0098] In addition, according to the training method of the complexity assessment system of the above embodiment of the present application, a complexity assessment system is trained, which includes: a feature extraction module for performing a first feature extraction operation on the acquired geometric data of the geometric model to obtain first feature data of the geometric model; a random forest module for performing a second feature extraction operation on the first feature data to obtain second feature data of the geometric model, and based on the second feature data, calculating the complexity category prediction value of the geometric model; a complexity assessment module for calculating the complexity assessment result of the geometric model based on the complexity category prediction value of the geometric model. For further details about the functions of the trained complexity assessment system, please refer to the description in the above training method embodiment, which will not be repeated here.

[0099] According to one embodiment of the present application, the complexity evaluation system trained in the above embodiment is applied to CAE software, so that the CAE software can perform performance simulation tests on geometric models with accurate complexity evaluation.

[0100] From the above description, it can be seen that applying the complexity evaluation system to CAE software enables CAE software to reasonably allocate computing resources for the performance simulation test of the geometric model based on the complexity evaluation results of the geometric model, thereby enabling CAE software to perform accurate and efficient performance simulation tests on the geometric model, thereby improving the production efficiency and quality of process products.

[0101] It should be noted that the geometric models processed by this application can include engineering part models such as automotive engine components and aerospace structural parts. Specifically, the above-mentioned embodiments proposed in this application can be used to improve the production efficiency and quality of automotive engine components and aerospace structural parts. The specific model structure processed is not limited to this. The technical solution of this application can be used to produce specific mechanical products, and no specific limitations are set here.

[0102] In order to verify the effectiveness of the above embodiments of the present application, the inventors conducted the following related experiments:

[0103] 1. Data Preparation

[0104] The dataset is divided into a training set and a test set, where the training set is 80% and the test set is 20%.

[0105] 2. Model Structure

[0106] ① Number of decision tree classifiers: Based on the complexity of the geometric model and the data scale, during the experiment, the number of decision tree classifiers in this application was set between 100 and 500, and the best experimental results were set to 100.

[0107] ② Decision tree classifier depth: The decision tree classifier depth of this application is set between 5 and 20 layers, and the best experimental results are set to 20.

[0108] It should be noted that the number and depth of decision tree classifiers can be adaptively adjusted according to experimental results. No specific restrictions are made here. The purpose is to prevent overfitting and underfitting and ensure that the decision tree classifier can fully learn the complex relationship between features and complexity.

[0109] 3. Training Process

[0110] ① Overfitting occurs during the training of the complexity evaluation system. This application adjusts the complexity of the complexity evaluation system by appropriately reducing the depth of the decision classification tree.

[0111] ② When the evaluation accuracy of the complexity evaluation system obtained through training is low, the complexity evaluation system is retrained and evaluated again by increasing the number of decision tree classifiers until the performance of the complexity evaluation system meets the expected requirements.

[0112] 4. Experimental Results

[0113] The performance of the trained complexity evaluation system is evaluated using the test set, and key evaluation indicators such as accuracy, precision, recall, and F1 value are calculated. The experimental results show that the construction method of this application can build a complexity evaluation system with better performance.

[0114] To sum up, compared with the existing geometric model complexity evaluation method, the above embodiment of the present application constructs an initial complexity evaluation system based on the random forest method, and uses the acquired geometric data of the geometric model to construct a data set to perform multiple iterative training on the initial complexity evaluation system until convergence, so as to obtain a complexity evaluation system that can evaluate the complexity of the geometric model more efficiently, objectively, accurately and universally.

[0115] In addition, this application proposes an application method of the complexity evaluation system, such as Figure 5 As shown, the method includes: T1, obtaining geometric data of a geometric model; T2, processing the geometric data using the complexity evaluation system of the above embodiment to obtain a complexity evaluation result of the geometric model.

[0116] It can be seen from the above description that the above embodiment of the present application inputs the geometric data of the geometric model into a trained and optimized complexity evaluation system, and the complexity evaluation system can output accurate complexity evaluation results of the geometric model.

[0117] To sum up, compared with the existing geometric model complexity evaluation method, which has the problem of being unable to evaluate the quality of the geometric model with high performance in terms of accuracy, objectivity, efficiency and versatility, the above-mentioned embodiment of the present application constructs an initial complexity evaluation system based on the random forest method, and uses the acquired geometric data of the geometric model to construct a data set to perform multiple iterative training on the initial complexity evaluation system until convergence, so as to obtain a complexity evaluation system that can evaluate the complexity of the geometric model more efficiently, objectively, accurately and universally, thereby improving the production efficiency and quality of process products.

[0118] Based on the inventive concept of the above embodiments, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the above embodiments when executing the computer program. Figure 6 Provide detailed explanation.

[0119] like Figure 6 As shown in FIG, the electronic device 100 of the present application is shown, which may specifically include a processor 110 and a memory 120. The memory 120 is coupled to the processor 110.

[0120] The processor 110 is used to control the operation of the electronic device. The processor 110 may also be referred to as a CPU (Central Processing Unit). The processor 110 may be an integrated circuit chip with signal processing capabilities. The processor 110 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. A general-purpose processor may be a microprocessor, or the processor 110 may be any conventional processor.

[0121] The memory 120 is used to store computer programs and can be RAM, ROM, or other types of storage terminals. Specifically, the memory 120 may include one or more computer-readable storage media, which may be non-transitory or transient. The memory 120 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage terminals and flash memory storage terminals. In some embodiments, the non-transitory computer-readable storage medium in the memory 120 is used to store at least one program code.

[0122] The processor 110 is configured to execute the computer program stored in the memory 120 to implement the methods described in the various method embodiments of the present application.

[0123] In some embodiments, the electronic device may further include a peripheral terminal interface 130 and at least one peripheral terminal. The processor 110, memory 120, and peripheral terminal interface 130 may be connected via a bus or signal lines. Each peripheral terminal may be connected to the peripheral terminal interface 130 via a bus, signal lines, or circuit boards. Specifically, the peripheral terminal includes at least one of a radio frequency circuit 140, a display screen 150, an audio circuit 160, and a power supply 170.

[0124] The peripheral terminal interface 130 can be used to connect at least one peripheral terminal related to I / O (Input / Output) to the processor 110 and the memory 120. In some embodiments, the processor 110, the memory 120, and the peripheral terminal interface 130 are integrated on the same chip or circuit board; in some other implementations, any one or two of the processor 110, the memory 120, and the peripheral terminal interface 130 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0125] The RF circuit 140 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 140 communicates with communication networks and other IoT devices via electromagnetic signals, and is therefore the communication circuitry of the electronic device. The RF circuit 140 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals into electrical signals. Optionally, the RF circuit 140 includes an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, an operator identification module card, and the like. The RF circuit 140 can communicate with other terminals via at least one wireless communication protocol. Such wireless communication protocols include, but are not limited to, the World Wide Web, metropolitan area networks, intranets, various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks, and / or WiFi (Wireless Fidelity) networks. In some embodiments, the RF circuit 140 may also include circuitry related to Near Field Communication (NFC), although this application does not limit this.

[0126] The display screen 150 is used to display a user interface (UI). The UI may include graphics, text, icons, videos, or any combination thereof. When the display screen 150 is a touch screen display, the display screen 150 is also capable of collecting touch signals on or above the surface of the display screen 150. The touch signals can be input as control signals to the processor 110 for processing. In this case, the display screen 150 can also be used to provide virtual buttons and / or virtual keyboards, also known as soft buttons and / or soft keyboards. In some embodiments, there can be one display screen 150, disposed on the front panel of the electronic device; in other embodiments, there can be at least two display screens 150, disposed on different surfaces of the electronic device or in a foldable design; in still other embodiments, the display screen 150 can be a flexible display, disposed on a curved or foldable surface of the electronic device. Furthermore, the display screen 150 can be configured as a non-rectangular irregular shape, i.e., a special-shaped screen. The display screen 150 can be made of materials such as LCD (Liquid Crystal Display) and OLED (Organic Light-Emitting Diode).

[0127] The audio circuit 160 may include a microphone and a speaker. The microphone is used to collect sound waves from the operator and the environment, and convert the sound waves into electrical signals and input them into the processor 110 for processing, or input them into the radio frequency circuit 140 to achieve voice communication. For the purpose of stereo acquisition or noise reduction, there can be multiple microphones, which are respectively set in different parts of the electronic device. The microphone can also be an array microphone or an omnidirectional acquisition microphone. The speaker is used to convert the electrical signal from the processor 110 or the radio frequency circuit 140 into sound waves. The speaker can be a traditional thin film speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can not only convert the electrical signal into sound waves audible to humans, but also convert the electrical signal into sound waves inaudible to humans for purposes such as ranging. In some embodiments, the audio circuit 160 may also include a headphone jack.

[0128] Power supply 170 is used to power various components in the electronic device. Power supply 170 can be AC power, DC power, a disposable battery, or a rechargeable battery. When power supply 170 includes a rechargeable battery, the rechargeable battery can be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery that is charged via a wired line, while a wireless rechargeable battery is a battery that is charged via a wireless coil. The rechargeable battery can also be used to support fast charging technology.

[0129] For a detailed description of the functions and execution processes of the various functional modules or components in the electronic device embodiments of the present application, reference can be made to the descriptions in the above-mentioned method embodiments of the present application, which will not be repeated here.

[0130] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the various embodiments of the electronic devices described above are merely illustrative. For example, the division of modules or units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0131] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0132] 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.

[0133] Based on the inventive concept of the above embodiments, the present application further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the steps of the method described in any of the above embodiments are described. Figure 7 The execution process of the above embodiment in a computer-readable storage medium is described.

[0134] like Figure 7 As shown, it shows the computer-readable storage medium of the present application. If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium 200. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions / computer programs for making an IoT device (which can be a personal computer, server, or network terminal, etc.) or a processor (processor) perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, as well as electronic terminals such as computers, mobile phones, laptops, tablet computers, cameras, etc. having the above-mentioned storage medium.

[0135] The description of the execution process of the program data in the computer-readable storage medium can refer to the description in the above-mentioned various method embodiments of the present application, and will not be repeated here.

[0136] The above description is merely an embodiment of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

[0137] Those skilled in the art will understand that in the above-mentioned method of the specific implementation method, the writing order of each step does not mean a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

Claims

1. A method for training a complexity evaluation system, wherein the complexity evaluation system is used to evaluate the complexity of a geometric model, characterized in that: And the method includes: Acquiring geometric data of the geometric model, preprocessing the geometric data, and constructing a data set based on the preprocessed geometric data; Construct an initial complexity assessment system based on the random forest method; The initial complexity evaluation system is trained multiple times by using the data set until convergence, so as to obtain the complexity evaluation system.

2. The training method of the complexity assessment system according to claim 1, characterized in that: The initial complexity evaluation system includes a feature extraction module, a random forest module and a complexity evaluation module, and the steps of constructing the initial complexity evaluation system based on the random forest method include: The feature extraction module is configured to perform a first feature extraction operation on the geometric data to obtain first feature data of the geometric model; The random forest module is configured to perform a second feature extraction operation on the first feature data to obtain second feature data of the geometric model, and calculate a complexity category prediction value of the geometric model based on the second feature data; The complexity evaluation module is configured to calculate a complexity evaluation result of the geometric model based on the complexity category prediction value of the geometric model.

3. The training method of the complexity assessment system according to claim 2, characterized in that: The initial complexity evaluation system further includes a standardization module, and the steps of constructing the initial complexity evaluation system based on the random forest method include: The standardization module is configured to perform standardization processing on the geometric data in the data set, and transmit the standardized results to the feature extraction module for processing.

4. The training method of the complexity assessment system according to claim 2, characterized in that: The geometric model includes faces, edges, and vertices. The first feature data of the geometric model includes basic topological feature data, curvature feature data, and geometric quantity feature data. The feature extraction module is configured to obtain the basic topological feature data, curvature feature data, and geometric quantity feature data of the geometric model in the following manner: Counting the number of faces, edges, and vertices in the geometric model, and using these data as basic topological feature data of the geometric model; Calculating the average curvature and Gaussian curvature of the surfaces in the geometric model, and counting the data and area ratios of the surfaces in the region where the curvature of the geometric model changes dramatically, and using the average curvature and Gaussian curvature of the surfaces, and the data and area ratios of the surfaces in the region where the curvature of the geometric model changes dramatically as curvature feature data of the geometric model; The number and geometric parameters of the holes, protrusions, and grooves in the geometric model are counted, and these data are used as geometric quantity feature data of the geometric model.

5. The training method of the complexity assessment system according to claim 2, characterized in that: The random forest module includes: multiple decision tree classifiers, and the step of constructing an initial complexity evaluation system based on the random forest method includes: The plurality of decision tree classifiers are respectively configured to perform a second feature extraction operation on the first feature data to obtain a plurality of second feature data of the geometric model, and calculate a plurality of complexity category prediction values of the geometric model based on the plurality of second feature data; The complexity evaluation module is also configured to calculate multiple initial complexity evaluation results of the geometric model based on a preset category threshold and multiple complexity category prediction values, and process the multiple initial complexity evaluation results according to preset rules to obtain the complexity evaluation result of the geometric model.

6. The training method of the complexity assessment system according to claim 1, characterized in that: The data set includes: a label corresponding to the geometric model, wherein the label indicates a true value of the complexity category of the geometric model, and the step of using the data set to iteratively train the initial complexity evaluation system multiple times until convergence to obtain the complexity evaluation system includes: During each iterative training process, the parameters of the complexity evaluation system are updated according to the following loss function: Where N is the number of geometric models in the dataset, y i represents the true value of the complexity category of the i-th geometric model in the dataset, represents the complexity category prediction value of the i-th geometric model in the dataset.

7. A complexity evaluation system trained using the complexity evaluation system training method according to claim 1, characterized in that: The complexity assessment system includes: a feature extraction module, configured to perform a first feature extraction operation on the acquired geometric data of the geometric model to obtain first feature data of the geometric model; a random forest module, configured to perform a second feature extraction operation on the first feature data to obtain second feature data of the geometric model, and calculate a complexity category prediction value of the geometric model based on the second feature data; The complexity evaluation module is used to calculate the complexity evaluation result of the geometric model based on the complexity category prediction value of the geometric model.

8. An application method of a complexity evaluation system, characterized in that: The method comprises: Obtaining geometric data of the geometric model; The complexity evaluation system according to claim 7 is used to process the geometric data to obtain a complexity evaluation result of the geometric model.

9. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 6 and 8 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 and 8 are implemented.

Images