Intelligent view recommendation method, device and equipment based on 3D modeling software

Abstract

The invention relates to the technical field of view recommendation, and discloses an intelligent view recommendation method, device and equipment based on 3D modeling software, and the method comprises the steps: obtaining historical modeling data in a historical modeling process, carrying out the preprocessing and feature extraction of the historical modeling data, and obtaining sample data; training a preset model by using the sample data to obtain a trained intelligent recommendation view model; and acquiring actual modeling data in an actual modeling process, inputting the actual modeling data into the intelligent recommendation view model, and performing view recommendation and dynamic adjustment. According to the intelligent view recommendation method based on the 3D modeling software provided by the invention, the intelligent view recommendation model is constructed, the most appropriate view angle is automatically recommended by utilizing the model in the actual modeling process, and the time and energy of the user are saved through an intelligent view switching mode, so that more accurate operation and making of a better decision are facilitated, and the user experience is improved. And the modeling efficiency and experience of the user in the 3D modeling process are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of view recommendation, and in particular to a method, device and equipment for intelligently recommending views based on 3D modeling software. Background Art

[0002] With the continuous development of computer technology, 3D modeling software is increasingly used in industrial design, mechanical manufacturing, architecture and other fields. However, in the 3D modeling process, traditional view display and switching usually require users to manually select or enter commands, which is very time-consuming in different modeling stages and complex scenarios. In addition, traditional view switching has the following problems:

[0003] (1) Users need to manually select a suitable perspective from multiple view options, which can be very tedious and time-consuming in different modeling stages and complex scenarios.

[0004] (2) Traditional view switching methods usually provide fixed view options and cannot be personalized according to the user's modeling habits and style.

[0005] (3) Editing or operating without a suitable viewing angle can easily lead to misoperation or the need for repeated adjustments, thereby reducing work efficiency.

[0006] (4) When dealing with complex models or scenes, traditional view switching methods may not provide enough perspective options to meet user needs, making it difficult for users to find the best perspective to observe and edit the model. Summary of the Invention

[0007] In view of this, the present invention provides a method, device and equipment for intelligently recommending views based on 3D modeling software to solve the above-mentioned problems existing in view display and switching during the 3D modeling process.

[0008] In a first aspect, the present invention provides a method for intelligently recommending views based on 3D modeling software, the method comprising:

[0009] Acquire historical modeling data during the historical modeling process, and perform preprocessing and feature extraction on the historical modeling data to obtain sample data;

[0010] Use sample data to train the preset model to obtain a trained intelligent recommendation view model;

[0011] Obtain the actual modeling data during the actual modeling process, and input the actual modeling data into the intelligent recommendation view model to perform view recommendation and dynamic adjustment.

[0012] The intelligent view recommendation method based on 3D modeling software provided by the present invention constructs an intelligent view recommendation model and uses the model to automatically recommend the most appropriate perspective during the actual modeling process. The intelligent view switching method saves users' time and energy, facilitates more precise operations and better decision-making, and improves users' modeling efficiency and experience in the 3D modeling process.

[0013] In an optional embodiment, the method further includes:

[0014] Obtain feedback data on view recommendations during the actual modeling process;

[0015] The feedback data is used to continuously optimize the intelligent recommendation view model to obtain an optimized intelligent recommendation view model.

[0016] The intelligent view recommendation method based on 3D modeling software provided by the present invention continuously optimizes the intelligent view recommendation model by collecting user feedback data on view recommendation results during the actual modeling process, improves model performance and view recommendation accuracy, enhances the functions and performance of 3D modeling software, and provides users with an efficient, convenient and personalized modeling experience.

[0017] In an optional implementation, the historical modeling data includes: view operation data, model state data, and context environment data. Acquiring the historical modeling data in the historical modeling process includes:

[0018] When starting modeling with 3D modeling software, the user's operations during the modeling process are continuously recorded as view operation data;

[0019] During the historical modeling process, the model status data at each moment is continuously obtained;

[0020] Obtain the project scene and environment information of the user's 3D model as contextual environment data in the historical modeling process.

[0021] The intelligent view recommendation method based on 3D modeling software provided by the present invention, by acquiring data related to view switching during the 3D modeling process, is conducive to accurately analyzing the main factors affecting the view, using the main factors to train the intelligent view recommendation model, and improving the accuracy and efficiency of view switching using the model.

[0022] In an optional embodiment, preprocessing and feature extraction are performed on historical modeling data to obtain sample data, including:

[0023] Clean, transform and normalize historical modeling data to obtain effective modeling data;

[0024] Use the preset convolutional neural network model to extract the texture features of the model surface and the context information of the view operation data in the effective modeling data;

[0025] The text description data in the effective modeling data is obtained, and the key information in the text description data is parsed using natural language processing. The texture features of the model surface, the context information of the view operation data and the key information in the text description data constitute the sample data.

[0026] The intelligent view recommendation method based on 3D modeling software provided by the present invention reduces errors in the data by cleaning the data, and can make the training model converge faster through normalization, which helps to save computing resources and time. Feature extraction can identify and retain the most critical information in the data set, while removing irrelevant or redundant features, simplifying the model and preventing overfitting.

[0027] In an optional implementation, the preset model is trained using sample data to obtain a trained intelligent recommendation view model, including:

[0028] Obtain the application scenario of the intelligent recommendation view and determine the preset model type based on the application scenario;

[0029] Select training sample data from the sample data according to the preset model type, and use the training sample data to train the preset model to obtain a trained recommendation model;

[0030] The recommendation model is evaluated and optimized using a cross-validation algorithm or a network search algorithm to obtain an intelligent recommendation view model.

[0031] In an optional embodiment, the preset model types include: support vector machine model, decision tree model, reinforcement learning model, and the preset model type is determined according to the application scenario, including:

[0032] If the application scenario is to recommend views based on 3D model types, select the support vector machine model as the preset model type;

[0033] If the application scenario is to recommend views based on user operation habits, select the decision tree model as the preset model type;

[0034] If the application scenario is to recommend views based on the user's repeated operation data, select the reinforcement learning model as the preset model type.

[0035] The intelligent view recommendation method based on 3D modeling software provided by the present invention selects different types of models for view recommendation according to different application scenarios, thereby improving view recommendation efficiency and computing resource utilization. Selecting appropriate models for specific application scenarios can significantly improve model performance. When processing structured data, support vector machines are more effective and easy to tune.

[0036] In an optional implementation, actual modeling data is input into the intelligent recommendation view model to perform view recommendation and dynamic adjustment, including:

[0037] Based on the application scenario of the smart recommendation view, select the smart recommendation view model that meets the application scenario;

[0038] Input the actual modeling data into the intelligent recommendation view model that meets the application scenario to make view recommendations;

[0039] Obtain changes in actual modeling data and determine changes in application scenarios based on changes in actual modeling data;

[0040] Dynamically adjust the type of the smart recommendation view model according to changes in the application scenario, and use the adjusted smart recommendation view model to perform view recommendations.

[0041] The intelligent view recommendation method based on 3D modeling software provided by the present invention dynamically adjusts the recommendation strategy according to changes in actual modeling data and application scenarios during the actual modeling process, automatically recommends the most appropriate viewing angle, reduces the time users spend looking for the appropriate viewing angle, improves modeling efficiency, enhances user experience, promotes precise operation, and assists decision-making.

[0042] In a second aspect, the present invention provides an intelligent view recommendation device based on 3D modeling software, the device comprising:

[0043] The sample data acquisition module is used to obtain historical modeling data in the historical modeling process, and preprocess and extract features of the historical modeling data to obtain sample data;

[0044] The model training module is used to train the preset model using sample data to obtain a trained intelligent recommendation view model;

[0045] The view recommendation module is used to obtain the actual modeling data in the actual modeling process, and input the actual modeling data into the intelligent recommendation view model to perform view recommendation and dynamic adjustment.

[0046] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0047] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method of the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 is a flowchart of a method for intelligently recommending views based on 3D modeling software according to an embodiment of the present invention;

[0050] Figure 2 is a flowchart of another method for intelligently recommending views based on 3D modeling software according to an embodiment of the present invention;

[0051] Figure 3 3D modeling software-based intelligent view recommendation method according to an embodiment of the present invention, a schematic diagram showing the comparison effect before and after view recommendation;

[0052] Figure 4 is a structural block diagram of an intelligent view recommendation device based on 3D modeling software according to an embodiment of the present invention;

[0053] Figure 5 Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0055] An embodiment of the present invention provides an intelligent view recommendation method based on 3D modeling software. By constructing an intelligent view recommendation model and using the model to automatically recommend the most appropriate perspective during the actual modeling process, the efficiency of view switching and user experience during the modeling process can be improved.

[0056] According to an embodiment of the present invention, an embodiment of a method for intelligently recommending views based on 3D modeling software is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0057] In this embodiment, a method for intelligently recommending views based on 3D modeling software is provided, which can be used in the above-mentioned computer system. Figure 1 FIG. 1 is a flow chart of a method for intelligently recommending views based on 3D modeling software according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0058] Step S101 : acquiring historical modeling data in a historical modeling process, and performing preprocessing and feature extraction on the historical modeling data to obtain sample data.

[0059] Specifically, historical modeling data includes a large amount of relevant data generated during the historical modeling process, including but not limited to user operation data and user behavior data. The core of the intelligent view recommendation method based on 3D modeling software provided in this embodiment is to learn and analyze historical modeling data. By leveraging algorithms such as deep learning, the 3D modeling software can identify which views provide the best viewing and operating experience for users at different modeling stages and user operations, and make view recommendations accordingly.

[0060] When users are working with 3D models with a lot of details and complex structures, such as mechanical parts, architectural models, or character sculptures, finding a suitable perspective to observe and edit a specific part can be time-consuming. The intelligent view recommendation method provided in this embodiment can automatically recommend a perspective that clearly displays the editing area based on the user's current editing operation (such as selection, movement, scaling, rotation, etc.) and the structural characteristics of the model, thereby improving editing efficiency and accuracy.

[0061] For historical modeling data, it is necessary to use neural networks such as CNN and RNN to extract the geometric and texture features of the 3D model and the contextual information of the user operation. For the text description related to the 3D model added by the user, natural language processing (NLP) technology can be used to parse the key information in the text description to obtain sample data. This is only an example and is not limited to this.

[0062] Step S102: Use sample data to train the preset model to obtain a trained intelligent recommendation view model.

[0063] Specifically, by combining the advantages of various machine learning algorithms, appropriate algorithms are selected in different application scenarios to build a preset model architecture, and the preset model is trained using sample data to obtain a trained intelligent recommendation view model.

[0064] For example: the Support Vector Machine (SVM) algorithm is mainly used to quickly recommend appropriate views based on existing operation data and historical modeling data; the decision tree algorithm is mainly used to recommend personalized views based on the user's operation habits and needs, thereby improving user convenience; the reinforcement learning algorithm (such as the Q-Learning algorithm) can learn based on the user's real-time view operation feedback, thereby more accurately grasping the user's view operation habits and operation processes. This is only an example, but not limited to this.

[0065] Step S103: obtaining actual modeling data in the actual modeling process, and inputting the actual modeling data into the intelligent recommendation view model to perform view recommendation and dynamic adjustment.

[0066] Specifically, when users use 3D modeling software for actual modeling, the intelligent recommendation view model will analyze the user's modeling operations in real time based on the actual modeling data, including but not limited to creation, editing, adjustment, and user interaction behaviors on the 3D modeling software interface.

[0067] The intelligent recommendation view model recommends and dynamically adjusts views based on real-time modeling data. For example, when a user is making fine edits to a model, such as adjusting the texture and details of the model surface, the intelligent recommendation view model may recommend a front view or a side view so that the user can see the editing effect more clearly. When a user is adjusting the overall layout of a model or performing scene design, the intelligent recommendation view model may recommend a top view or a perspective view so that the user can grasp the structure and proportion of the model or scene from a more macro perspective. When a user performs some complex operations, such as rotating, scaling, and moving a model, the intelligent recommendation view model may recommend a view that can simultaneously display multiple key surfaces of the model so that the user can observe and operate more comprehensively. This is just an example, but not limited to this.

[0068] The intelligent view recommendation method based on 3D modeling software provided in this embodiment constructs an intelligent view recommendation model and uses the model to automatically recommend the most appropriate perspective during the actual modeling process. The intelligent view switching method saves users' time and energy, facilitates more precise operations and better decision-making, and improves users' modeling efficiency and experience in the 3D modeling process.

[0069] In some optional embodiments, the method further comprises:

[0070] Step S104: obtaining feedback data on view recommendations during the actual modeling process.

[0071] Specifically, during the actual modeling process, users will check whether the recommended views meet their needs. If the view recommendation is not suitable, the user will manually adjust it. The user's operation data during the manual adjustment of the view can be automatically obtained as feedback data for the view recommendation. In addition, a convenient feedback entrance is set in the 3D modeling software interface. When the user feels that the recommended perspective does not meet his or her expectations, for example, the recommended perspective for displaying the building courtyard landscape does not highlight the user's carefully designed gardening part, the user can submit opinions through the feedback entrance, and the opinions submitted through the feedback entrance can also be used as feedback data.

[0072] Step S105 : continuously optimizing the intelligent recommendation view model using the feedback data to obtain an optimized intelligent recommendation view model.

[0073] Specifically, after the software operator collects feedback data on view recommendations, it uses the feedback data to continuously learn and optimize the intelligent recommendation view model, adjust the weights of feature extraction, improve algorithm parameters, etc., so that when encountering similar scenarios again in the future, it can more accurately recommend perspectives that meet user needs, thereby improving recommendation accuracy and user satisfaction.

[0074] The intelligent view recommendation method based on 3D modeling software provided in this embodiment continuously optimizes the intelligent view recommendation model by collecting user feedback data on view recommendation results during the actual modeling process, improves model performance and view recommendation accuracy, enhances the functions and performance of 3D modeling software, and provides users with an efficient, convenient and personalized modeling experience.

[0075] In this embodiment, a method for intelligently recommending views based on 3D modeling software is provided, which can be used in the above-mentioned computer system. Figure 2 FIG. 1 is a flow chart of a method for intelligently recommending views based on 3D modeling software according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0076] Step S201 : acquiring historical modeling data in the historical modeling process, and performing preprocessing and feature extraction on the historical modeling data to obtain sample data.

[0077] Specifically, the historical modeling data includes: view operation data, model state data, and context environment data. The above step S201 includes:

[0078] S2011, when starting modeling using 3D modeling software, the user's operations during the modeling process are continuously recorded as view operation data.

[0079] Specifically, when the user starts modeling in the software, various operations performed by the user when creating the building model will be continuously recorded as view operation data, including the creation, editing, and adjustment of the 3D model, as well as the user's interactive behavior on the interface, such as the user rotating the view of each face of the building model. This is only an example, but not limited to this.

[0080] S2012, during the historical modeling process, the model status data at each moment is continuously obtained.

[0081] Specifically, the model status data includes relevant information about the stage of the current modeling process, such as relevant status information about whether the current model is in the overall framework construction stage or the detail decoration stage, which is only an example and not limited to this.

[0082] S2013, obtaining the project scene and environment information of the 3D model created by the user as contextual environment data in the historical modeling process.

[0083] Specifically, the project scene can be contextual environmental data such as whether the project for which the user creates a 3D model is an interior design scene or an outdoor landscape scene (for example, whether there is a current need for lighting simulation, etc.), and the environmental information can be contextual environmental data such as the complexity of the model, the currently edited part, and the user's modeling habits. This is only an example, but not limited to this.

[0084] The intelligent view recommendation method based on 3D modeling software provided in this embodiment, by acquiring data related to view switching during the 3D modeling process, is conducive to accurately analyzing the main factors affecting the view, using the main factors to train the intelligent view recommendation model, and improving the accuracy and efficiency of view switching using the model.

[0085] Step S2014: clean, convert, and normalize the historical modeling data to obtain valid modeling data.

[0086] Specifically, the acquired raw modeling data needs to be preprocessed to obtain valid modeling data. For example, some operation data may contain duplicate or erroneous records. For example, if the same operation is recorded multiple times due to a brief software freeze, these redundant data need to be cleaned up. Model status data indicators in different formats are uniformly converted into standardized formats to facilitate subsequent processing. The angle data in view operations is normalized and limited to a specific reasonable range so that subsequent algorithms can be better used.

[0087] Step S2015: Use a preset convolutional neural network model to extract the texture features of the model surface in the effective modeling data and the context information of the view operation data.

[0088] Specifically, a convolutional neural network (CNN) is used to perform convolution processing on the effective modeling data to extract texture features on the surface of the 3D building model, such as the geometric texture features such as the wall brick texture of the building facade and the reflective texture of the glass curtain wall. Contextual information such as the order in which users perform viewing operations on different building parts is analyzed, and the contextual information of the viewing operation data is used to determine the areas that users focus on and their operating habits.

[0089] Step S2016, obtain text description data from the valid modeling data, and use natural language processing to parse out key information in the text description data. The texture features of the model surface, the context information of the view operation data and the key information in the text description data constitute the sample data.

[0090] Specifically, if the user adds some text descriptions of the building's functions and styles in the software (such as "building a European classical-style villa"), NLP technology will be used to parse out key information such as "European classical" and "villa" to assist subsequent models in understanding user needs.

[0091] The sample data obtained after preprocessing the historical modeling data includes: texture features of the model surface, context information of the view operation data and key information in the text description data.

[0092] The intelligent view recommendation method based on 3D modeling software provided in this embodiment reduces errors in the data by cleaning the data, and can achieve faster convergence when training the model through normalization, which helps to save computing resources and time. Feature extraction can identify and retain the most critical information in the data set, while removing irrelevant or redundant features, simplifying the model and preventing overfitting.

[0093] Step S202: Use sample data to train the preset model to obtain a trained intelligent recommendation view model.

[0094] Specifically, the above step S202 includes:

[0095] Step S2021: Obtain the application scenario of the intelligent recommendation view and determine the preset model type according to the application scenario.

[0096] Specifically, the application scenarios of intelligent recommended views are determined based on sample data. For example, when a user makes detailed edits to a part of a 3D building model, repeated operations may be required. The Q-learning algorithm can be selected as the preset model type. The Q-learning algorithm can learn based on the user's real-time view operation feedback, thereby more accurately grasping the user's view operation habits and operation processes.

[0097] In addition to the Q-learning algorithm, you can also select a support vector machine (SVM) or a decision tree algorithm as the preset model type. The SVM algorithm is mainly used to quickly recommend suitable views based on existing operation data and historical records; the decision tree algorithm is mainly used to recommend personalized views based on the operating habits and needs of the target user, thereby improving the user experience of the target user. This is only an example, but not limited to this.

[0098] Step S2022: Select training sample data from the sample data according to the preset model type, and use the training sample data to train the preset model to obtain a trained recommendation model.

[0099] Specifically, based on the preset model type, training sample data used in the preset model training process is selected from the sample data for model training to obtain a trained recommendation model. For example, if a user repeatedly zooms in on a wall in a 3D building model, the preset model of the Q-learning algorithm will use the user's operation data from the sample data as training sample data, without requiring model state data. This is for example only and is not intended to be limiting.

[0100] In a specific embodiment, in different scenarios, different preset models are selected and different historical modeling data are used for training the models. The following are historical modeling data corresponding to different scenarios, which are only examples and are not limited thereto:

[0101] In scenario a, when the user is grading the quality of a historical 3D building model, they select a preset model corresponding to the SVM algorithm. The historical modeling data includes:

[0102] (1) Model status data:

[0103] Accurate dimensional measurements: This includes the length, width, height, diameter, and wall thickness of each part of a 3D building model. These measurements are the basis for determining whether the model conforms to the design blueprint. For example, deviations in the critical dimensions of a custom mechanical part directly affect subsequent assembly and use.

[0104] Surface microscopic features: The surface roughness value obtained by the roughness measuring instrument directly reflects the smoothness of the model surface and affects the product's appearance, touch, and fit with other components.

[0105] Internal structure analysis: Information such as internal porosity and filler distribution obtained through 3D scanning or slice analysis is crucial for the integrity and uniformity of the internal structure of some functional 3D printed products, such as aerospace parts.

[0106] (2) View operation data:

[0107] Measurement process records: Detailed records of the measurement steps of tools such as calipers and micrometers, including measurement positions and measurement sequences, ensure the accuracy and repeatability of the measurements for subsequent tracing and analysis.

[0108] Appearance visual inspection: Record the angle of observation while rotating around the model and the duration of observation in key areas to accurately capture possible defects on the model surface, such as bubbles and dents.

[0109] In scenario c, when troubleshooting a 3D mechanical assembly process, the user selects a preset model corresponding to the decision tree algorithm. The historical modeling data includes:

[0110] (1) Model status data:

[0111] Dimension deviation analysis: data such as shaft diameter deviation, hole inner diameter deviation, keyway position deviation, etc. These dimensional deviations are often the direct cause of assembly failure.

[0112] Surface defect statistics: length, depth, location of scratches, direction and size of cracks, as well as the degree and scope of surface corrosion, etc. These defects will seriously affect the performance and assembly quality of parts.

[0113] Form and position tolerance considerations: Form and position tolerance data such as roundness, cylindricity, flatness, and perpendicularity play a key role in ensuring the accuracy and stability of mechanical assembly.

[0114] (2) Contextual data:

[0115] Environmental factor monitoring: temperature, humidity, air pressure and other data of the assembly environment, because some metals are prone to rust in humid environments, and temperature changes will cause parts to expand and contract, thus affecting assembly accuracy.

[0116] Vibration condition record: Vibration conditions at the assembly site. Excessive vibration may cause parts to shift or be damaged during the assembly process, affecting assembly quality.

[0117] (3) Text description data:

[0118] Assembly process guide: Detailed assembly sequence, assembly tool usage, tightening torque and other requirements are important bases for judging whether the assembly process is correct.

[0119] Fault Case Library: The phenomena, cause analysis, and solutions of past assembly failures provide valuable reference and experience for current troubleshooting.

[0120] In scenario e, when a user searches for a specific exhibit in a 3D virtual museum, they select a preset model corresponding to the Q-learning algorithm. The historical modeling data includes:

[0121] (1) View operation data:

[0122] Navigation operation records: Users’ forward, backward, and turn operations in the virtual museum, as well as records of the use of the map navigation function, including clicks on map locations, zooming in and out of the map, etc., reflecting the user’s movement in the space.

[0123] Perspective switching operation: The operation of rotating the perspective to observe the surrounding environment, including the angle and speed of perspective rotation, and the time of observation in different directions, which helps users obtain more environmental information.

[0124] (2) Contextual data:

[0125] Museum layout information: the distribution of exhibition halls, the display location of exhibits, the direction of aisles, etc. This information constitutes the spatial framework for users to find exhibits.

[0126] Crowd flow information: The movement paths and gathering areas of other virtual visitors, which may indicate popular exhibits or guide users to find the direction of the target exhibits.

[0127] (3) Text description data:

[0128] Exhibit description information: the name of the exhibit, its age, related introduction, etc. These textual information help users identify and locate the target exhibits.

[0129] In-museum prompt information: such as text prompts on signboards, broadcast notifications, etc., providing users with clues to find exhibits.

[0130] Step S2023: Use a cross-validation algorithm or a network search algorithm to evaluate and optimize the recommendation model to obtain an intelligent recommendation view model.

[0131] Specifically, through cross-validation, the training sample data is divided into multiple subsets, and different subsets are used in turn as validation sets to evaluate the accuracy of the recommendation model. A grid search method is used to traverse different parameter combinations, such as the learning rate and number of neurons in the neural network, to optimize the model parameters and evaluate its performance, continuously improving the model's recommendation effect. This process continues until the recommendation model converges and a smart recommendation view model is obtained.

[0132] In some optional implementations, the preset model types include: support vector machine model, decision tree model, reinforcement learning model, and step S2021 determines the preset model type according to the application scenario, including:

[0133] If the application scenario is to recommend views based on 3D model types, the support vector machine model is selected as the preset model type.

[0134] Specifically, the 3D modeling software has accumulated a large amount of operational data and historical records of multiple users in the past when creating different types of building models (such as residential buildings, commercial buildings, etc.). Using the SVM algorithm, based on the existing massive historical data, it can quickly recommend suitable views to new users who are creating residential models and are in the basic wall construction stage. For example, it can show whether the wall layout is reasonable from conventional angles such as the front and side, helping new users to efficiently confirm the accuracy of the basic part of the model, thereby improving modeling efficiency and user experience.

[0135] If the application scenario is to recommend views based on user operating habits, select the decision tree model as the preset model type.

[0136] Specifically, based on the learning of user operation behaviors, the intelligent recommendation view model can provide users with personalized view recommendations to adapt to the modeling styles and needs of different users.

[0137] For professional architectural designers with specific operating habits, such as those who prioritize overall layout over detailed interior design, the decision tree algorithm provides personalized view recommendations based on the target user's past operating habits and project requirements, such as "focusing on space efficiency." For example, when modeling a new commercial space, views that showcase spatial divisions and pedestrian flow are prioritized, improving user convenience and allowing them to quickly focus on key design aspects.

[0138] If the application scenario is to recommend views based on the user's repeated operation data, select the reinforcement learning model as the preset model type.

[0139] Specifically, when users adjust the shape of a building's roof in real time, rotating and zooming the view, the Q-Learning algorithm can learn from these real-time viewing feedback. For example, if a user frequently zooms in to view the drainage structure details at a certain angle on the roof, the algorithm can understand that the user is paying close attention to this detail. Subsequently, it can more accurately recommend appropriate close-up, multi-angle views of the roof details based on the user's viewing habits and operation processes.

[0140] The intelligent view recommendation method based on 3D modeling software provided in this embodiment selects different types of models for view recommendation according to different application scenarios, thereby improving view recommendation efficiency and computing resource utilization. Selecting appropriate models for specific application scenarios can significantly improve model performance. When processing structured data, support vector machines are more effective and easier to tune.

[0141] Step S203: Acquire actual modeling data during the actual modeling process, and input the actual modeling data into the intelligent recommendation view model to perform view recommendation and dynamic adjustment.

[0142] Specifically, the above step S203 includes:

[0143] Step S2031: Select an intelligent recommendation view model that meets the application scenario based on the application scenario of the intelligent recommendation view.

[0144] Specifically, for example, if a user creates a residential 3D building model, the intelligent recommendation view model of the SVM algorithm is first used to recommend reasonable views of the residential 3D building model to the user.

[0145] Step S2032: Input the actual modeling data into the intelligent recommendation view model that meets the application scenario to perform view recommendation.

[0146] Specifically, the operation data, model status data and contextual environment data of the user in the process of creating a 3D building model are input into the intelligent recommendation view model to recommend a suitable perspective for the user.

[0147] Step S2033: Acquire changes in actual modeling data, and determine changes in application scenarios based on the changes in the actual modeling data.

[0148] Specifically, during the user modeling process, changes in actual modeling data are captured, and changes in application scenarios are determined based on these changes. For example, when working on a 3D residential building model, a user frequently zooms in to view the drainage structure details at a certain angle on the roof. This indicates that the application scenario is to recommend views based on the user's repeated operation data. In this case, the type of intelligent recommendation view needs to be changed to the Q-learning algorithm.

[0149] In a specific embodiment, a user builds a 3D model of a car engine. During the modeling process, the user frequently observes the assembly relationship of parts. The comparison effect before and after the intelligent view recommendation method provided by this embodiment is as follows: Figure 3 As shown, the left picture is the effect before view recommendation, which is a global view and not the best perspective for observing the assembly relationship of parts. The right picture is the effect after view recommendation, which is a partially enlarged view and the best perspective for observing the assembly relationship of parts.

[0150] Step S2034: dynamically adjust the type of the intelligent recommendation view model according to changes in the application scenario, and use the adjusted intelligent recommendation view model to perform view recommendation.

[0151] Specifically, as the modeling process progresses and user operations change, the recommended views will also be dynamically adjusted to always maintain relevance to the user's current task. As the user continues to improve the building model, such as when starting to add door and window components, data such as the user's new view operations at this time (such as frequently switching to the interior to check whether the door and window positions are reasonable) and the new state of the model (the overall style coordination after the doors and windows are added) are collected in real time. This real-time data is input into the trained model for inference, and perspectives that can better show the matching effect of doors and windows with indoor and outdoor spaces are recommended in a timely manner, such as views that show the lighting and aesthetics of doors and windows from different indoor and outdoor angles. At the same time, if the user temporarily changes the project requirements, such as changing from an ordinary residence to a smart home display model house, the context environment changes, and the model will dynamically adjust the recommendation strategy based on this change, recommending more relevant perspectives such as the reserved space for smart device installation and line layout.

[0152] The intelligent view recommendation method based on 3D modeling software provided in this embodiment dynamically adjusts the recommendation strategy according to changes in actual modeling data and application scenarios during the actual modeling process, automatically recommends the most appropriate perspective, reduces the time users spend looking for the right perspective, improves modeling efficiency, enhances user experience, promotes precise operation, and assists decision-making.

[0153] In this embodiment, a device for intelligently recommending views based on 3D modeling software is also provided. The device is used to implement the above-mentioned embodiments and preferred implementations, and the details that have been described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0154] This embodiment provides an intelligent view recommendation device based on 3D modeling software, such as Figure 4 Shown, including:

[0155] The sample data acquisition module 401 is used to acquire historical modeling data in the historical modeling process, and perform preprocessing and feature extraction on the historical modeling data to obtain sample data.

[0156] The model training module 402 is used to train the preset model using sample data to obtain a trained intelligent recommendation view model.

[0157] The view recommendation module 403 is used to obtain actual modeling data in the actual modeling process, and input the actual modeling data into the intelligent recommendation view model to perform view recommendation and dynamic adjustment.

[0158] In some optional implementations, the sample data acquisition module 401 includes:

[0159] The operation data recording unit is used to continuously record the user's operations during the modeling process as view operation data when starting modeling using 3D modeling software.

[0160] The model status data acquisition unit is used to continuously obtain the model status data at each moment during the historical modeling process.

[0161] The environmental data acquisition unit is used to obtain the project scene and environmental information of the 3D model created by the user as contextual environmental data in the historical modeling process.

[0162] The data preprocessing unit is used to clean, convert and normalize the historical modeling data to obtain effective modeling data.

[0163] The feature extraction unit is used to extract the texture features of the model surface and the context information of the view operation data in the effective modeling data using a preset convolutional neural network model.

[0164] The sample data determination unit is used to obtain text description data from valid modeling data and use natural language processing to parse out key information in the text description data. The texture features of the model surface, the context information of the view operation data and the key information in the text description data constitute the sample data.

[0165] In some optional implementations, the model training module 402 includes:

[0166] The model type determination unit is used to obtain the application scenario of the intelligent recommendation view and determine the preset model type according to the application scenario.

[0167] The model training unit is used to select training sample data from the sample data according to the preset model type, and use the training sample data to train the preset model to obtain a trained recommendation model.

[0168] The model optimization unit is used to evaluate and optimize the recommendation model using a cross-validation algorithm or a network search algorithm to obtain an intelligent recommendation view model.

[0169] In some optional implementations, the view recommendation module 403 includes:

[0170] The recommendation model selection unit is used to select an intelligent recommendation view model that meets the application scenario based on the application scenario of the intelligent recommendation view.

[0171] The view recommendation unit is used to input actual modeling data into the intelligent recommendation view model that meets the application scenario to make view recommendations.

[0172] The modeling process data change acquisition unit is used to obtain changes in actual modeling data and determine changes in application scenarios based on changes in actual modeling data.

[0173] The model dynamic adjustment unit is used to dynamically adjust the type of the intelligent recommendation view model according to changes in the application scenario, and use the adjusted intelligent recommendation view model to perform view recommendation.

[0174] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0175] The intelligent recommendation view device based on 3D modeling software in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0176] The embodiment of the present invention also provides a computer device having the above Figure 4 The intelligent recommendation view device based on 3D modeling software is shown.

[0177] See also Figure 5 , Figure 5 is a structural diagram of a computer device provided by an optional embodiment of the present invention, such as Figure 5 As shown, the computer device includes: one or more processors 10, memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 10 is taken as an example.

[0178] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0179] The memory 20 stores instructions that can be executed by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0180] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0181] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0182] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0183] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0184] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. An intelligent view recommendation method based on 3D modeling software, characterized in that: The method comprises: Acquiring historical modeling data during the historical modeling process, and performing preprocessing and feature extraction on the historical modeling data to obtain sample data; Using the sample data to train a preset model to obtain a trained intelligent recommendation view model; The actual modeling data in the actual modeling process is obtained, and the actual modeling data is input into the intelligent recommendation view model to perform view recommendation and dynamic adjustment.

2. The method according to claim 1, characterized in that The method further comprises: Obtain feedback data on view recommendations during the actual modeling process; The intelligent recommendation view model is continuously optimized using the feedback data to obtain an optimized intelligent recommendation view model.

3. The method according to claim 1 or 2, characterized in that The historical modeling data includes: view operation data, model state data, and context environment data. Acquiring the historical modeling data during the historical modeling process includes: When starting modeling with 3D modeling software, the user's operations during the modeling process are continuously recorded as view operation data; During the historical modeling process, the model status data at each moment is continuously obtained; Obtain the project scene and environment information of the user's 3D model as contextual environment data in the historical modeling process.

4. The method according to claim 1, wherein Preprocessing and feature extraction are performed on the historical modeling data to obtain sample data, including: Cleaning, converting and normalizing the historical modeling data to obtain effective modeling data; Using a preset convolutional neural network model to extract texture features of the model surface and context information of the view operation data in the effective modeling data; The text description data in the effective modeling data is obtained, and the key information in the text description data is parsed using natural language processing. The texture features of the model surface, the context information of the view operation data and the key information in the text description data constitute the sample data.

5. The method according to claim 1, wherein The preset model is trained using the sample data to obtain a trained intelligent recommendation view model, including: Obtaining an application scenario of the intelligent recommendation view, and determining a preset model type according to the application scenario; Selecting training sample data from the sample data according to the preset model type, and using the training sample data to train the preset model to obtain a trained recommendation model; The recommendation model is evaluated and optimized using a cross-validation algorithm or a network search algorithm to obtain an intelligent recommendation view model.

6. The method according to claim 5, characterized in that The preset model types include: support vector machine model, decision tree model, reinforcement learning model, and the preset model type is determined according to the application scenario, including: If the application scenario is to recommend views based on 3D model types, the support vector machine model is selected as the preset model type; If the application scenario is to recommend views based on user operating habits, the default model type is a decision tree model; If the application scenario is to recommend views based on the user's repeated operation data, the preset model type selects the reinforcement learning model.

7. The method according to claim 6, characterized in that Inputting the actual modeling data into the intelligent recommendation view model to perform view recommendation and dynamic adjustment, including: According to the application scenario of the intelligent recommendation view, select an intelligent recommendation view model that meets the application scenario; Inputting the actual modeling data into an intelligent recommendation view model that meets the application scenario to perform view recommendation; Acquiring changes in actual modeling data, and determining changes in application scenarios based on the changes in the actual modeling data; Dynamically adjust the type of the smart recommendation view model according to changes in the application scenario, and use the adjusted smart recommendation view model to perform view recommendations.

8. An intelligent view recommendation device based on 3D modeling software, characterized in that: The device comprises: A sample data acquisition module is used to acquire historical modeling data in the historical modeling process, and preprocess and extract features from the historical modeling data to obtain sample data; A model training module is used to train a preset model using the sample data to obtain a trained intelligent recommendation view model; The view recommendation module is used to obtain actual modeling data in the actual modeling process, and input the actual modeling data into the intelligent recommendation view model to perform view recommendation and dynamic adjustment.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method according to any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

Images