Slice model support scheme prediction method and system based on neural network

Abstract

The application discloses a slice model support scheme prediction method and system based on a neural network, and the method comprises the following steps: acquiring slice data of three-dimensional model data to be subjected to 3D printing; identifying all support members corresponding to the three-dimensional model data according to the slice data; calculating the weakness degree of each support member corresponding to the support member according to the member parameter corresponding to the support member; and predicting a support optimization scheme corresponding to the three-dimensional model data according to the weakness degree of the support member based on a prediction neural network model; and the support optimization scheme is used for limiting the member change parameter corresponding to at least one support member. It can be seen that the application can realize accurate support structure optimization based on the member characteristics and the weakness degree, improve the stability and the printing success rate of the 3D printing model, and reduce the printing failure risk caused by insufficient support.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for predicting schemes based on a slice model supported by a neural network. Background Technology

[0002] With the widespread application of 3D printing technology in high-precision manufacturing, enterprises and users are increasingly focusing on improving model stability and printing success rates by optimizing support structures. Existing technologies typically acquire slice data from 3D models, identify support components using basic geometric analysis or fixed rules, and generate support structures based on standard parameter settings to support the printing process. However, existing solutions lack dynamic calculations of component weakness and predictive optimization using neural network models, making it difficult to accurately assess support requirements and generate targeted solutions. Commonly used static support design strategies cannot adapt to complex model structures, leading to insufficient or excessive support, which can easily cause printing failures or material waste, limiting the stability and efficiency of the 3D printing process. Therefore, existing technologies have shortcomings that urgently need to be addressed. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for predicting support schemes for sliced ​​models based on neural networks, which can achieve accurate support structure optimization based on component characteristics and weakness, improve the stability and printing success rate of 3D printed models, and reduce the risk of printing failure due to insufficient support.

[0004] To address the aforementioned technical problems, the first aspect of this invention discloses a method for predicting a slice model based on a neural network, the method comprising:

[0005] Obtain slice data of the 3D model data to be 3D printed;

[0006] Based on the slice data, all supporting components corresponding to the three-dimensional model data are identified;

[0007] Based on the component parameters corresponding to the supporting components, the degree of weakness corresponding to each supporting component is calculated;

[0008] Based on a predictive neural network model, a support optimization scheme corresponding to the three-dimensional model data is predicted according to the weakness of the support component; the support optimization scheme is used to limit at least one component change parameter corresponding to the support component.

[0009] As an optional implementation, in the first aspect of the invention, identifying all supporting components corresponding to the three-dimensional model data based on the slice data includes:

[0010] The slice layers at any two adjacent positions in the slice data are input into the trained support connection prediction neural network to obtain the predicted probability of the two slice layers belonging to the support connection site.

[0011] Calculate the image difference between the two slice layers;

[0012] Calculate the product of the predicted probability and the image difference to obtain the support connectivity parameters corresponding to the two slice layers;

[0013] Based on the support connection parameters, all support components corresponding to the three-dimensional model data are identified.

[0014] As an optional implementation, in the first aspect of the present invention, the support connection prediction neural network is an RNN network architecture, which is trained using a training dataset including multiple sets of training continuous slice layers and corresponding labels indicating whether they belong to support connection locations.

[0015] As an optional implementation, in the first aspect of the invention, identifying all support components corresponding to the three-dimensional model data based on the support connection parameters includes:

[0016] For each slice layer, calculate the average value of all the support connection parameters corresponding to that slice layer to obtain the possible support parameters corresponding to that slice layer;

[0017] All slice layers whose possible support parameters are greater than a preset first parameter threshold are selected and identified as multiple support slice layers.

[0018] Based on the support slice layer, all support components corresponding to the three-dimensional model data are identified.

[0019] As an optional implementation, in the first aspect of the invention, identifying all support components corresponding to the three-dimensional model data based on the support slice layer includes:

[0020] For each of the aforementioned support slice layers, based on an image segmentation algorithm, multiple component cross-sectional envelope patterns belonging to different components in the support slice layer are segmented out;

[0021] An image containing multiple cross-sectional envelope patterns of the components is input into a trained support component prediction neural network to obtain the probability of each cross-sectional envelope pattern of the components belonging to the support component.

[0022] The components whose cross-sectional envelope patterns have a probability greater than a preset first probability threshold are identified as support components in the three-dimensional model data.

[0023] As an optional implementation, in the first aspect of the present invention, the supporting component prediction neural network is an RNN network architecture, which is trained using a training dataset that includes multiple training component cross-sectional images and corresponding annotations indicating whether they belong to supporting components.

[0024] As an optional implementation, in the first aspect of the invention, calculating the weakness of each support member based on the member parameters corresponding to the support member includes:

[0025] Determine the component parameters corresponding to each of the aforementioned support components; the component parameters include the component surface area, component volume, and at least one associated component with a connection relationship;

[0026] For each of the aforementioned support members, the reciprocal of the weighted sum of the member's surface area and volume is calculated to obtain the member's stability parameters.

[0027] Calculate the average value of the component stability parameters of all the associated components corresponding to the support component to obtain the associated load-bearing parameters;

[0028] The sum of the component's stability parameter and the associated load-bearing parameter is calculated to obtain the corresponding weakness of the support component.

[0029] As an optional implementation, in the first aspect of the present invention, the step of predicting the support optimization scheme corresponding to the three-dimensional model data based on the predictive neural network model and according to the weakness of the support component includes:

[0030] Based on a clustering algorithm, at least one set of support components is identified; the set of support components includes multiple support components whose relative positions are less than a preset distance threshold and whose weakness is greater than a preset degree threshold.

[0031] The component parameters and weakness of all support components in each set of support components are input into a trained parameter change prediction neural network to obtain the component parameter change scheme corresponding to each set of support components; the parameter change prediction neural network is trained on a training dataset that includes multiple training component parameters and corresponding weakness labeling and parameter optimization scheme labeling;

[0032] The component parameter variation schemes corresponding to all the aforementioned support component sets are determined as the support optimization schemes corresponding to the three-dimensional model data.

[0033] A second aspect of this invention discloses a slice model-supported prediction system based on a neural network, the system comprising:

[0034] The acquisition module is used to acquire slice data of the 3D model data to be 3D printed;

[0035] The identification module is used to identify all supporting components corresponding to the three-dimensional model data based on the slice data;

[0036] The calculation module is used to calculate the degree of weakness of each of the supporting components based on the component parameters corresponding to the supporting components.

[0037] The prediction module is used to predict the support optimization scheme corresponding to the three-dimensional model data based on the prediction neural network model and the weakness of the support component; the support optimization scheme is used to limit at least one component change parameter corresponding to the support component.

[0038] As an optional implementation, in a second aspect of the invention, the specific method by which the identification module identifies all supporting components corresponding to the three-dimensional model data based on the slice data includes:

[0039] The slice layers at any two adjacent positions in the slice data are input into the trained support connection prediction neural network to obtain the predicted probability of the two slice layers belonging to the support connection site.

[0040] Calculate the image difference between the two slice layers;

[0041] Calculate the product of the predicted probability and the image difference to obtain the support connectivity parameters corresponding to the two slice layers;

[0042] Based on the support connection parameters, all support components corresponding to the three-dimensional model data are identified.

[0043] As an optional implementation, in a second aspect of the invention, the support connection prediction neural network is an RNN network architecture, which is trained using a training dataset that includes multiple sets of training continuous slice layers and corresponding labels indicating whether they belong to support connection locations.

[0044] As an optional implementation, in a second aspect of the invention, the identification module identifies the specific method by which it identifies all supporting components corresponding to the three-dimensional model data based on the supporting connection parameters, including:

[0045] For each slice layer, calculate the average value of all the support connection parameters corresponding to that slice layer to obtain the possible support parameters corresponding to that slice layer;

[0046] All slice layers whose possible support parameters are greater than a preset first parameter threshold are selected and identified as multiple support slice layers.

[0047] Based on the support slice layer, all support components corresponding to the three-dimensional model data are identified.

[0048] As an optional implementation, in a second aspect of the invention, the specific method by which the identification module identifies all supporting components corresponding to the three-dimensional model data based on the supporting slice layer includes:

[0049] For each of the aforementioned support slice layers, based on an image segmentation algorithm, multiple component cross-sectional envelope patterns belonging to different components in the support slice layer are segmented out;

[0050] An image containing multiple cross-sectional envelope patterns of the components is input into a trained support component prediction neural network to obtain the probability of each cross-sectional envelope pattern of the components belonging to the support component.

[0051] The components whose cross-sectional envelope patterns have a probability greater than a preset first probability threshold are identified as support components in the three-dimensional model data.

[0052] As an optional implementation, in a second aspect of the invention, the supporting component prediction neural network is an RNN network architecture, which is trained using a training dataset that includes multiple cross-sectional images of training components and corresponding annotations indicating whether they belong to supporting components.

[0053] As an optional implementation, in a second aspect of the invention, the specific method by which the calculation module calculates the weakness of each support member based on the member parameters corresponding to the support member includes:

[0054] Determine the component parameters corresponding to each of the aforementioned support components; the component parameters include the component surface area, component volume, and at least one associated component with a connection relationship;

[0055] For each of the aforementioned support members, the reciprocal of the weighted sum of the member's surface area and volume is calculated to obtain the member's stability parameters.

[0056] Calculate the average value of the component stability parameters of all the associated components corresponding to the support component to obtain the associated load-bearing parameters;

[0057] The sum of the component's stability parameter and the associated load-bearing parameter is calculated to obtain the corresponding weakness of the support component.

[0058] As an optional implementation, in a second aspect of the invention, the prediction module, based on a predictive neural network model, predicts the specific method of the support optimization scheme corresponding to the three-dimensional model data according to the weakness of the support component, including:

[0059] Based on a clustering algorithm, at least one set of support components is identified; the set of support components includes multiple support components whose relative positions are less than a preset distance threshold and whose weakness is greater than a preset degree threshold.

[0060] The component parameters and weakness of all support components in each set of support components are input into a trained parameter change prediction neural network to obtain the component parameter change scheme corresponding to each set of support components; the parameter change prediction neural network is trained on a training dataset that includes multiple training component parameters and corresponding weakness labeling and parameter optimization scheme labeling;

[0061] The component parameter variation schemes corresponding to all the aforementioned support component sets are determined as the support optimization schemes corresponding to the three-dimensional model data.

[0062] A third aspect of this invention discloses another neural network-based slice model support scheme prediction system, the system comprising:

[0063] Memory containing executable program code;

[0064] A processor coupled to the memory;

[0065] The processor calls the executable program code stored in the memory to execute some or all of the steps in the neural network-based slice model support scheme prediction method disclosed in the first aspect of the present invention.

[0066] The fourth aspect of the present invention discloses a computer storage medium storing computer instructions, which, when invoked, are used to execute some or all of the steps in the neural network-based slice model support scheme prediction method disclosed in the first aspect of the present invention.

[0067] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0068] This invention acquires slice data of 3D printed three-dimensional model data and identifies support components. It calculates the weakness of the support components based on component parameters and generates a support optimization scheme based on a predictive neural network model. This enables precise support structure optimization based on component characteristics and weakness, improving the stability and printing success rate of 3D printed models and reducing the risk of printing failure due to insufficient support. Attached Figure Description

[0069] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0070] Figure 1 This is a flowchart illustrating a slice model-based prediction method for predicting schemes based on neural networks, as disclosed in an embodiment of the present invention.

[0071] Figure 2 This is a schematic diagram of the structure of a slice model-supported prediction system based on a neural network disclosed in an embodiment of the present invention.

[0072] Figure 3 This is a schematic diagram of another neural network-based slice model support scheme prediction system disclosed in an embodiment of the present invention. Detailed Implementation

[0073] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0074] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0075] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0076] This invention discloses a method and system for predicting support schemes for sliced ​​models based on neural networks. By acquiring sliced ​​data of a 3D printed model and identifying support components, and calculating the weakness of these components based on their parameters, an optimized support scheme is generated based on a predictive neural network model. This enables precise optimization of the support structure based on component characteristics and weakness, improving the stability and printing success rate of the 3D printed model, and reducing the risk of printing failure due to insufficient support. Detailed explanations follow.

[0077] Example 1

[0078] Please see Figure 1 , Figure 1 This is a flowchart illustrating a scheme prediction method based on a neural network-supported slice model, as disclosed in an embodiment of the present invention. Figure 1 The described neural network-based slice model-supported prediction method can be applied to data processing systems / data processing equipment / data processing servers (including local processing servers or cloud processing servers). For example... Figure 1 As shown, the neural network-based slice model-supported scheme prediction method may include the following operations:

[0079] 101. Obtain slice data of the 3D model data to be 3D printed.

[0080] 102. Based on the slice data, identify all supporting components corresponding to the 3D model data.

[0081] 103. Calculate the degree of weakness of each supporting component based on the component parameters corresponding to the supporting components.

[0082] 104. Based on the predictive neural network model, the support optimization scheme corresponding to the three-dimensional model data is predicted according to the weakness of the support components.

[0083] Optionally, the support optimization scheme is used to define the component variation parameters corresponding to at least one support component.

[0084] Optionally, the 3D model data can be an STL file, an OBJ file, or an AMF file; this invention does not impose any limitations.

[0085] Optionally, the slice data may include layer thickness data, contour path data, or filling structure data, and the present invention does not limit it.

[0086] Optionally, the acquisition of the slice data can be achieved based on slicing software, real-time processing, or cloud computing; this invention does not limit the scope of the invention.

[0087] Optionally, the support member can be a single support column, a grid support structure, or a tree-like support structure; the present invention does not limit this.

[0088] Optionally, the identification process of the support component can be implemented based on a neural network model, geometric analysis, or rule matching algorithm, and the present invention does not limit it.

[0089] Optionally, the component parameters may include the component surface area, component volume, component height, or material density, which are not limited in this invention.

[0090] Optionally, the degree of weakness can be a risk score, a stability index, or a failure probability; this invention does not limit this.

[0091] Optionally, the component change parameters may include component size adjustment, position movement, or material change parameters, which are not limited in this invention.

[0092] Optionally, the support optimization scheme may include optimizing the number of supports, adjusting the support layout, or enhancing the support strength; this invention does not limit the scope of the scheme.

[0093] As can be seen, the above-described embodiments of the invention acquire slice data of 3D printed three-dimensional model data and identify support components, calculate the degree of weakness by combining component parameters, and generate support optimization schemes based on predictive neural network models. This enables precise support structure optimization based on component characteristics and degree of weakness, improves the stability and printing success rate of 3D printed models, and reduces the risk of printing failure due to insufficient support.

[0094] As an optional embodiment, the step above, identifying all supporting components corresponding to the 3D model data based on the slice data, includes:

[0095] Input any two adjacent slice layers in the slice data into the trained support connection prediction neural network to obtain the predicted probability that the two slice layers belong to the support connection site.

[0096] Calculate the image difference between the two slice layers;

[0097] The product of the predicted probability and the image difference is calculated to obtain the support connectivity parameters corresponding to the two slice layers;

[0098] Based on the support connection parameters, all support components corresponding to the 3D model data are identified.

[0099] Optionally, the supporting connection prediction neural network can be a deep learning classification model, a convolutional neural network, or a probabilistic prediction model; this invention does not limit the type of neural network.

[0100] Optionally, the predicted probability can be a binary probability, a multi-class probability, or a confidence score; this invention does not impose any limitations.

[0101] Optionally, the training process of the supporting connection prediction neural network can be implemented based on historical slice data, simulated data, or labeled data, and the present invention does not limit it.

[0102] Optionally, the image difference can be pixel difference, structural similarity index, or feature vector distance; this invention does not limit the specific difference.

[0103] Optionally, the calculation of the image difference can be based on image processing algorithms, feature extraction, or statistical analysis, and this invention does not limit it.

[0104] As can be seen, through the above optional embodiments, by inputting adjacent slice layers into a trained support connection prediction neural network to obtain prediction probabilities, combining image differences to calculate support connection parameters and identifying support components accordingly, accurate support component identification based on slice layer correlation and image features can be achieved, thereby improving the accuracy of 3D printing support structure analysis and reducing the risk of misjudgment of support components.

[0105] As an optional embodiment, in the above steps, the support connection prediction neural network is an RNN network architecture, which is trained using a training dataset that includes multiple sets of training continuous slice layers and corresponding labels indicating whether they belong to support connection locations.

[0106] As can be seen, the model details of the support connection prediction neural network are defined through the above optional embodiments, so as to accurately predict the support connection position in the future, assist in the accurate optimization of the support structure based on the component characteristics and weakness, improve the stability of the 3D printing model and the printing success rate, and reduce the risk of printing failure due to insufficient support.

[0107] As an optional embodiment, the step above, identifying all supporting components corresponding to the 3D model data based on the supporting connection parameters, includes:

[0108] For each slice layer, calculate the average value of all support connection parameters corresponding to that slice layer to obtain the possible support parameters corresponding to that slice layer;

[0109] Filter out all slice layers whose possible support parameters are greater than the preset first parameter threshold, and determine them as multiple support slice layers;

[0110] Based on the support slice layer, all support components corresponding to the 3D model data are identified.

[0111] Optionally, the first parameter threshold can be a fixed threshold, a dynamic threshold, or a threshold adjusted based on model complexity; this invention does not impose any limitations.

[0112] As can be seen, through the above optional embodiments, the possible support parameters are obtained by calculating the average support connection parameters between each slice layer and adjacent slice layers. Slice layers with parameters exceeding the threshold are selected as support slice layers and support components are identified, thereby achieving accurate support slice layer selection based on connection parameter analysis, improving the efficiency and accuracy of 3D printing support component identification, and reducing the risk of structural identification errors.

[0113] As an optional embodiment, the step described above, identifying all support components corresponding to the 3D model data based on the support slice layer, includes:

[0114] For each support slice layer, based on the image segmentation algorithm, the cross-sectional envelope patterns of multiple components belonging to different components in the support slice layer are segmented out;

[0115] An image containing the envelope patterns of multiple component cross sections is input into a trained support component prediction neural network to obtain the probability of each component belonging to a support component corresponding to its cross section envelope pattern.

[0116] The components whose cross-sectional envelope patterns have a probability greater than a preset first probability threshold are identified as support components in the 3D model data.

[0117] Optionally, the image segmentation algorithm can be the U-Net algorithm, the Mask R-CNN algorithm, or the DeepLab algorithm; this invention does not limit the algorithm.

[0118] Optionally, the cross-sectional envelope of the component can be a two-dimensional contour, a three-dimensional projection, or a simplified geometric figure; the present invention does not limit this.

[0119] Optionally, the predictive neural network of the supporting component can be a convolutional neural network, a deep learning classification model, or a probabilistic prediction model; this invention does not impose any limitations.

[0120] Optionally, the component probability can be a binary probability, a multi-class probability, or a confidence score; this invention does not impose any limitations.

[0121] Optionally, the training process of the support component prediction neural network can be implemented based on historical slice data, simulated component data, or labeled data, and the present invention does not limit it.

[0122] Optionally, the first probability threshold can be a fixed threshold, a dynamic threshold, or a threshold adjusted based on the printing task; the present invention does not limit this.

[0123] As can be seen, through the above optional embodiments, by segmenting the cross-sectional envelope pattern of the component in the support slice layer based on the image segmentation algorithm, inputting the trained support component prediction neural network to obtain the component probability and determine the support component, the accurate support component identification based on image segmentation and probability prediction is achieved, thereby improving the accuracy and reliability of 3D printed support structure analysis and reducing the risk of component misjudgment.

[0124] As an optional embodiment, in the above steps, the supporting component prediction neural network is an RNN network architecture, which is trained using a training dataset that includes multiple training component cross-sectional images and corresponding annotations indicating whether they belong to supporting components.

[0125] As can be seen, the model details of the neural network for predicting support components are defined through the above optional embodiments, so as to accurately predict the support components in the future, assist in the precise optimization of the support structure based on the characteristics and weakness of the components, improve the stability of the 3D printing model and the printing success rate, and reduce the risk of printing failure due to insufficient support.

[0126] As an optional embodiment, the step above, calculating the weakness of each supporting member based on the member parameters, includes:

[0127] Determine the component parameters corresponding to each supporting component; optionally, the component parameters include the component surface area, component volume, and at least one associated component with a connection relationship.

[0128] For each supporting member, calculate the reciprocal of the weighted sum of the member's surface area and volume to obtain the member's stability parameters;

[0129] Calculate the average value of the stability parameters of all associated components corresponding to the supporting component to obtain the associated load-bearing parameters;

[0130] The sum of the structural stability parameters and the associated load-bearing parameters of the component is calculated to obtain the degree of weakness of the supporting component.

[0131] As can be seen, through the above optional embodiments, the component stability parameters are obtained by calculating the weighted sum of the surface area and volume of the support component, and the weakness is calculated by combining the average stability parameters of the associated components. This enables accurate weakness assessment based on component parameters and relationships, improves the targeting and printing stability of 3D printed support structure optimization, and reduces the risk of support failure.

[0132] As an optional embodiment, the above steps, including predicting the support optimization scheme corresponding to the three-dimensional model data based on the predictive neural network model and the corresponding weakness of the support component, include:

[0133] Based on a clustering algorithm, at least one set of support components is identified; optionally, the set of support components includes multiple support components whose relative positions are less than a preset distance threshold and whose weakness is greater than a preset degree threshold.

[0134] The component parameters and weakness of all support components in each support component set are input into a trained parameter change prediction neural network to obtain the component parameter change scheme corresponding to each support component set; optionally, the parameter change prediction neural network is trained using a training dataset that includes multiple training component parameters and corresponding weakness labeling and parameter optimization scheme labeling.

[0135] The parameter variation schemes corresponding to all support component sets are determined as the support optimization schemes corresponding to the 3D model data.

[0136] Optionally, the clustering algorithm can be the K-means algorithm, the hierarchical clustering algorithm, or the DBSCAN algorithm; this invention does not limit the algorithm.

[0137] Optionally, the distance threshold can be a fixed distance, a dynamic distance, or a distance adjusted based on the model size; this invention does not impose any limitations.

[0138] Optionally, the neural network for predicting parameter changes can be a regression neural network, a classification neural network, or an adversarial generative network model; this invention does not limit the specific model.

[0139] As can be seen, through the above optional embodiments, by identifying the set of support components whose positional distance and weakness meet the threshold based on the clustering algorithm, and combining the parameter change prediction neural network to generate component parameter change schemes and integrate them into support optimization schemes, accurate support optimization based on clustering and prediction is achieved, thereby improving the structural stability and printing quality of 3D printing models and reducing the risk of printing defects caused by insufficient support.

[0140] Example 2

[0141] Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of a slice model-supported prediction system based on a neural network, as disclosed in an embodiment of the present invention. Figure 2 The described neural network-based slice model-supported prediction system can be applied to data processing systems / data processing equipment / data processing servers (including local processing servers or cloud processing servers). For example... Figure 2 As shown, the neural network-based slice model-supported prediction system may include:

[0142] The acquisition module 201 is used to acquire slice data of the three-dimensional model data to be 3D printed.

[0143] The identification module 202 is used to identify all supporting components corresponding to the three-dimensional model data based on the slice data.

[0144] The calculation module 203 is used to calculate the degree of weakness of each support component based on the component parameters corresponding to the support components.

[0145] The prediction module 204 is used to predict the support optimization scheme corresponding to the three-dimensional model data based on the prediction neural network model and the corresponding weakness of the support component.

[0146] Optionally, the support optimization scheme is used to define the component variation parameters corresponding to at least one support component.

[0147] As can be seen, the above-described embodiments of the invention acquire slice data of 3D printed three-dimensional model data and identify support components, calculate the degree of weakness by combining component parameters, and generate support optimization schemes based on predictive neural network models. This enables precise support structure optimization based on component characteristics and degree of weakness, improves the stability and printing success rate of 3D printed models, and reduces the risk of printing failure due to insufficient support.

[0148] As an optional embodiment, the identification module identifies the specific method by which it identifies all supporting components corresponding to the 3D model data based on the slice data, including:

[0149] Input any two adjacent slice layers in the slice data into the trained support connection prediction neural network to obtain the predicted probability that the two slice layers belong to the support connection site.

[0150] Calculate the image difference between the two slice layers;

[0151] The product of the predicted probability and the image difference is calculated to obtain the support connectivity parameters corresponding to the two slice layers;

[0152] Based on the support connection parameters, all support components corresponding to the 3D model data are identified.

[0153] As can be seen, through the above optional embodiments, by inputting adjacent slice layers into a trained support connection prediction neural network to obtain prediction probabilities, combining image differences to calculate support connection parameters and identifying support components accordingly, accurate support component identification based on slice layer correlation and image features can be achieved, thereby improving the accuracy of 3D printing support structure analysis and reducing the risk of misjudgment of support components.

[0154] As an optional embodiment, the support connection prediction neural network is an RNN network architecture, which is trained on a training dataset that includes multiple sets of training continuous slice layers and corresponding labels indicating whether they belong to support connection locations.

[0155] As can be seen, the model details of the support connection prediction neural network are defined through the above optional embodiments, so as to accurately predict the support connection position in the future, assist in the accurate optimization of the support structure based on the component characteristics and weakness, improve the stability of the 3D printing model and the printing success rate, and reduce the risk of printing failure due to insufficient support.

[0156] As an optional embodiment, the identification module identifies the specific method by which it identifies all supporting components corresponding to the 3D model data based on the supporting connection parameters, including:

[0157] For each slice layer, calculate the average value of all support connection parameters corresponding to that slice layer to obtain the possible support parameters corresponding to that slice layer;

[0158] Filter out all slice layers whose possible support parameters are greater than the preset first parameter threshold, and determine them as multiple support slice layers;

[0159] Based on the support slice layer, all support components corresponding to the 3D model data are identified.

[0160] As can be seen, through the above optional embodiments, the possible support parameters are obtained by calculating the average support connection parameters between each slice layer and adjacent slice layers. Slice layers with parameters exceeding the threshold are selected as support slice layers and support components are identified, thereby achieving accurate support slice layer selection based on connection parameter analysis, improving the efficiency and accuracy of 3D printing support component identification, and reducing the risk of structural identification errors.

[0161] As an optional embodiment, the identification module identifies the specific method by which it identifies all supporting components corresponding to the 3D model data based on the supporting slice layer, including:

[0162] For each support slice layer, based on the image segmentation algorithm, the cross-sectional envelope patterns of multiple components belonging to different components in the support slice layer are segmented out;

[0163] An image containing the envelope patterns of multiple component cross sections is input into a trained support component prediction neural network to obtain the probability of each component belonging to a support component corresponding to its cross section envelope pattern.

[0164] The components whose cross-sectional envelope patterns have a probability greater than a preset first probability threshold are identified as support components in the 3D model data.

[0165] As can be seen, through the above optional embodiments, by segmenting the cross-sectional envelope pattern of the component in the support slice layer based on the image segmentation algorithm, inputting the trained support component prediction neural network to obtain the component probability and determine the support component, the accurate support component identification based on image segmentation and probability prediction is achieved, thereby improving the accuracy and reliability of 3D printed support structure analysis and reducing the risk of component misjudgment.

[0166] As an optional embodiment, the support component prediction neural network is an RNN network architecture, which is trained on a training dataset that includes multiple cross-sectional images of training components and corresponding annotations indicating whether they belong to support components.

[0167] As can be seen, the model details of the neural network for predicting support components are defined through the above optional embodiments, so as to accurately predict the support components in the future, assist in the precise optimization of the support structure based on the characteristics and weakness of the components, improve the stability of the 3D printing model and the printing success rate, and reduce the risk of printing failure due to insufficient support.

[0168] As an optional embodiment, the calculation module calculates the specific method for determining the weakness of each supporting member based on the member parameters, including:

[0169] Determine the component parameters corresponding to each supporting component; optionally, the component parameters include the component surface area, component volume, and at least one associated component with a connection relationship.

[0170] For each supporting member, calculate the reciprocal of the weighted sum of the member's surface area and volume to obtain the member's stability parameters;

[0171] Calculate the average value of the stability parameters of all associated components corresponding to the supporting component to obtain the associated load-bearing parameters;

[0172] The sum of the structural stability parameters and the associated load-bearing parameters of the component is calculated to obtain the degree of weakness of the supporting component.

[0173] As can be seen, through the above optional embodiments, the component stability parameters are obtained by calculating the weighted sum of the surface area and volume of the support component, and the weakness is calculated by combining the average stability parameters of the associated components. This enables accurate weakness assessment based on component parameters and relationships, improves the targeting and printing stability of 3D printed support structure optimization, and reduces the risk of support failure.

[0174] As an optional embodiment, the prediction module, based on a predictive neural network model, predicts the specific method of the support optimization scheme corresponding to the 3D model data according to the weakness of the supporting components, including:

[0175] Based on a clustering algorithm, at least one set of support components is identified; optionally, the set of support components includes multiple support components whose relative positions are less than a preset distance threshold and whose weakness is greater than a preset degree threshold.

[0176] The component parameters and weakness of all support components in each support component set are input into a trained parameter change prediction neural network to obtain the component parameter change scheme corresponding to each support component set; optionally, the parameter change prediction neural network is trained using a training dataset that includes multiple training component parameters and corresponding weakness labeling and parameter optimization scheme labeling.

[0177] The parameter variation schemes corresponding to all support component sets are determined as the support optimization schemes corresponding to the 3D model data.

[0178] As can be seen, through the above optional embodiments, by identifying the set of support components whose positional distance and weakness meet the threshold based on the clustering algorithm, and combining the parameter change prediction neural network to generate component parameter change schemes and integrate them into support optimization schemes, accurate support optimization based on clustering and prediction is achieved, thereby improving the structural stability and printing quality of 3D printing models and reducing the risk of printing defects caused by insufficient support.

[0179] Example 3

[0180] Please see Figure 3 , Figure 3 This is another neural network-based slice model support scheme prediction system disclosed in the embodiments of the present invention. Figure 3 The described neural network-based slice model-supported prediction system is applied in data processing systems / data processing equipment / data processing servers (wherein, the server includes local processing servers or cloud processing servers). For example... Figure 3 As shown, the neural network-based slice model-supported prediction system may include:

[0181] Memory 301 storing executable program code;

[0182] Processor 302 coupled to memory 301;

[0183] The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the neural network-based slice model support scheme prediction method described in Embodiment 1.

[0184] Example 4

[0185] This invention discloses a computer read storage medium that stores a computer program for electronic data interchange, wherein the computer program causes a computer to execute the steps of the neural network-based slice model support scheme prediction method described in Embodiment 1.

[0186] Example 5

[0187] This invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps of the neural network-based slice model support scheme prediction method described in Embodiment 1.

[0188] The foregoing has described specific embodiments of this specification; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0189] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0190] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0191] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0192] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0193] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0194] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0195] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0196] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0197] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0198] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0199] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0200] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0201] Finally, it should be noted that the method and system for predicting a slice model based on a neural network disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting a slicing scheme based on a neural network-based model, characterized in that, The method includes: Obtain slice data of the 3D model data to be 3D printed; Based on the slice data, all supporting components corresponding to the 3D model data are identified, including: The slice layers at any two adjacent positions in the slice data are input into the trained support connection prediction neural network to obtain the predicted probability of the two slice layers belonging to the support connection site. Calculate the image difference between the two slice layers; Calculate the product of the predicted probability and the image difference to obtain the support connectivity parameters corresponding to the two slice layers; For each slice layer, calculate the average value of all the support connection parameters corresponding to that slice layer to obtain the possible support parameters corresponding to that slice layer; All slice layers whose possible support parameters are greater than a preset first parameter threshold are selected and identified as multiple support slice layers. For each of the aforementioned support slice layers, based on an image segmentation algorithm, multiple component cross-sectional envelope patterns belonging to different components in the support slice layer are segmented out; An image containing multiple cross-sectional envelope patterns of the components is input into a trained support component prediction neural network to obtain the probability of each cross-sectional envelope pattern of the components belonging to the support component. The components whose cross-sectional envelope patterns of all the components with a probability greater than a preset first probability threshold are identified as support components in the three-dimensional model data. Based on the component parameters corresponding to the supporting components, the degree of weakness corresponding to each supporting component is calculated; Based on a predictive neural network model, a support optimization scheme corresponding to the three-dimensional model data is predicted according to the weakness of the support component; the support optimization scheme is used to limit at least one component change parameter corresponding to the support component.

2. The method for predicting a slicing model based on a neural network according to claim 1, characterized in that, The support connection prediction neural network is an RNN network architecture, which is trained using a training dataset that includes multiple sets of training continuous slice layers and corresponding labels indicating whether they belong to support connection locations.

3. The method for predicting a slicing model based on a neural network according to claim 1, characterized in that, The supporting component prediction neural network is an RNN network architecture, which is trained using a training dataset that includes multiple cross-sectional images of training components and corresponding annotations indicating whether they belong to supporting components.

4. The method for predicting a slicing model based on a neural network according to claim 1, characterized in that, The step of calculating the weakness of each support member based on the member parameters of the support member includes: Determine the component parameters corresponding to each of the aforementioned support components; the component parameters include the component surface area, component volume, and at least one associated component with a connection relationship; For each of the aforementioned support members, the reciprocal of the weighted sum of the member's surface area and volume is calculated to obtain the member's stability parameters. Calculate the average value of the component stability parameters of all the associated components corresponding to the support component to obtain the associated load-bearing parameters; The sum of the component's stability parameter and the associated load-bearing parameter is calculated to obtain the corresponding weakness of the support component.

5. The method for predicting a slicing model based on a neural network according to claim 1, characterized in that, The method of predicting the support optimization scheme corresponding to the three-dimensional model data based on the predictive neural network model and the corresponding weakness of the support component includes: Based on a clustering algorithm, at least one set of support components is identified; the set of support components includes multiple support components whose relative positions are less than a preset distance threshold and whose weakness is greater than a preset degree threshold. The component parameters and weakness of all support components in each set of support components are input into a trained parameter change prediction neural network to obtain the component parameter change scheme corresponding to each set of support components; the parameter change prediction neural network is trained on a training dataset that includes multiple training component parameters and corresponding weakness labeling and parameter optimization scheme labeling; The component parameter variation schemes corresponding to all the aforementioned support component sets are determined as the support optimization schemes corresponding to the three-dimensional model data.

6. A scheme prediction system based on a slice model supported by a neural network, characterized in that, The system executes the neural network-based slice model support scheme prediction method as described in any one of claims 1-5, the system comprising: The acquisition module is used to acquire slice data of the 3D model data to be 3D printed; The identification module is used to identify all supporting components corresponding to the three-dimensional model data based on the slice data; The calculation module is used to calculate the degree of weakness of each of the supporting components based on the component parameters corresponding to the supporting components. The prediction module is used to predict the support optimization scheme corresponding to the three-dimensional model data based on the prediction neural network model and the weakness of the support component; the support optimization scheme is used to limit at least one component change parameter corresponding to the support component.

7. A scheme prediction system based on a slice model supported by a neural network, characterized in that, The system includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the neural network-based slice model support scheme prediction method as described in any one of claims 1-5.

Images