Paper folding design method and system based on diffusion model and physical information neural network

Abstract

The invention provides an origami design method and system based on a diffusion model and a physical information neural network, and belongs to the technical field of origami structures, and the method comprises the steps: obtaining point cloud data of a target three-dimensional origami shape and corresponding text description information; according to the point cloud data and the text description information, generating a two-dimensional crease pattern by using a graph diffusion model, and ensuring that the crease pattern conforms to a specified three-dimensional shape by using conditional coding; and optimizing the two-dimensional fold pattern by adopting a physical information neural network to obtain a final two-dimensional origami graph. According to the method, the diffusion model and the physical constraint neural network are combined, high-precision modeling of the crease pattern is achieved, it is ensured that the generated pattern can be folded geometrically and physically, and therefore the design efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of origami structures, and in particular relates to an origami design method and system based on a diffusion model and a physical information neural network. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Origami inverse design is a computational technique used to derive the corresponding two-dimensional crease pattern from a known three-dimensional folded shape. This technique has a wide range of applications in fields such as materials science, robotics, and biomedicine. By inversely reasoning about two-dimensional crease patterns, foldable structures with complex geometries can be designed, thereby achieving the goals of saving space and improving functionality. For example, origami structures can be used to manufacture retractable satellite components or micro medical devices that can be unfolded from a compact state when needed. However, current inverse origami design methods have certain challenges.

[0004] Conventional origami reverse design methods mainly rely on geometric rules and manual experience. Designers gradually unfold the three-dimensional model into a plane and manually deduce its crease pattern. However, this method is very time-consuming when dealing with complex shapes, and it is difficult to ensure that the generated crease pattern meets the actual physical constraints. This experience-based design method usually relies on existing origami theories such as Kawasaki's Theorem and Maekawa's Theorem. These theories ensure that origami patterns can be folded physically flat, but they cannot automatically deduce the two-dimensional crease pattern of arbitrary three-dimensional structures.

[0005] In recent years, researchers have begun exploring inverse design methods that use computer algorithms to automatically generate origami patterns. Some studies have addressed the design of origami structures using geometric methods, such as continuous geometric models. While these methods can generate crease patterns that conform to geometric constraints to a certain extent, they often exhibit limitations when generating complex and diverse origami structures.

[0006] With advances in deep learning technology, some research has attempted to use generative models to automate origami reverse design. For example, some methods use deep generative models to generate two-dimensional crease patterns that conform to geometric and physical constraints. However, these models often require complex post-processing steps to ensure the foldability of the generated patterns, and the generated patterns still lack physical feasibility and accuracy. Summary of the Invention

[0007] In order to overcome the shortcomings of the above-mentioned existing technologies, the present invention proposes an origami design method and system based on a diffusion model and a physical information neural network, which is used to derive the corresponding two-dimensional crease pattern from a known three-dimensional folding structure. Combining the diffusion model and the physical constraint neural network, it can generate diversified two-dimensional crease patterns that meet the physical folding requirements, solving the shortcomings of traditional methods in generating crease patterns in complexity and physical feasibility.

[0008] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: In a first aspect, an origami design method based on a diffusion model and a physical information neural network is disclosed, comprising: Obtain point cloud data and corresponding text description information of the target three-dimensional origami shape; Generate a two-dimensional crease pattern using a graphic diffusion model based on the point cloud data and text description information, and use conditional coding to ensure that the crease pattern conforms to a specified three-dimensional shape; The two-dimensional crease pattern is optimized using a physical information neural network to obtain a final two-dimensional origami figure.

[0009] In the second aspect, an origami design system based on a diffusion model and a physical information neural network is disclosed, including: a data acquisition module configured to: acquire point cloud data of a target three-dimensional origami shape and corresponding text description information; a preliminary crease generation module configured to: generate a two-dimensional crease pattern using a graphic diffusion model based on the point cloud data and text description information, and ensure that the crease pattern conforms to a specified three-dimensional shape using conditional coding; The origami generation module is configured to optimize the two-dimensional crease pattern using a physical information neural network to obtain a final two-dimensional origami diagram.

[0010] In a third aspect, an electronic device is disclosed, including a memory and a processor, and computer instructions stored in the memory and running on the processor. When the computer instructions are run by the processor, the steps of the origami design method based on the diffusion model and physical information neural network are completed.

[0011] In a fourth aspect, a computer-readable storage medium is disclosed for storing computer instructions. When the computer instructions are executed by a processor, the steps of the origami design method based on the diffusion model and the physical information neural network are completed.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. Diversity and physical feasibility of generated crease patterns: The present invention combines a diffusion model with a physically constrained neural network to generate diverse two-dimensional crease patterns that meet physical folding requirements, addressing the shortcomings of traditional methods in terms of complexity and physical feasibility in generating crease patterns.

[0013] 2. Enhanced design controllability: By combining multimodal conditional embedding of three-dimensional point clouds and natural language descriptions, the present invention achieves precise control over the generated crease patterns, which can meet specific origami design requirements.

[0014] 3. Improve computational efficiency: Compared with traditional geometric derivation and optimization methods, this invention significantly improves the efficiency of generating crease patterns and reduces the design time of complex shapes by combining the diffusion generation model and neural network.

[0015] 4. Combination of physical and geometric optimization: Physically constrained neural networks ensure that the generated crease patterns are not only geometrically accurate but also physically foldable, addressing the shortcomings of origami patterns in existing technologies in terms of physical feasibility.

[0016] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0018] Figure 1 This is a flow chart of the origami design method based on the diffusion model and physical information neural network described in Example 1 of the present invention.

[0019] Figure 2 This is the 3D point cloud origami, 2D crease diagram, and 3D crease shape paper crane sample diagram described in Example 1 of the present invention.

[0020] Figure 3 This is a schematic diagram of the data expansion preprocessing results described in Example 1 of the present invention.

[0021] Figure 4 This is a model diagram of the origami design method based on the diffusion model and physical information neural network described in Example 1 of the present invention.

[0022] Figure 5 This is a schematic diagram of the graph-based conditional diffusion model described in Example 1 of the present invention.

[0023] Figure 6 This is a schematic diagram of the physical information neural network described in Example 1 of the present invention.

[0024] Figure 7 This is a schematic diagram of the two-dimensional origami diagram generated as described in Example 1 of the present invention.

[0025] Figure 8 This is a set of example diagrams of the results of the origami reverse design method based on the diffusion model and physical information neural network in Example 1 of the present invention. DETAILED DESCRIPTION

[0026] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0027] It should be noted that the terms used herein are for describing particular embodiments only and are not intended to limit the exemplary embodiments according to the present invention.

[0028] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0029] Example 1 In one or more embodiments, an origami design method based on a diffusion model and a physical information neural network is disclosed, such as Figure 1 As shown, the following steps are included: Step S1: obtaining point cloud data of a target three-dimensional origami shape and corresponding text description information; During the training phase, this embodiment uses publicly available 3D classic origami models (from OrigamiSimulator) and custom generated crease pattern data for experiments. The 3D data comes from online origami simulators (including FreeformOrigami and OrigamiSimulator). These data contain the 3D point cloud of the origami model and the corresponding 2D crease pattern. These data are used as the dataset for model training. The dataset includes origami models of different shapes, such as cranes, ships, frogs, etc., covering various folding complexities. Figure 2 Shown are 3D point cloud origami, 2D crease map, and 3D crease shape paper crane sample images.

[0030] The original 3D origami data and 2D origami graph structure are expanded to enrich the training data and enhance the generalization ability of the model. The expansion methods include standard data augmentation techniques, including rotation and flipping, noise perturbation and scale transformation. The 3D point cloud and crease pattern are rotated and mirrored along different axes to increase the directional diversity of the data. A small amount of Gaussian noise is added to the point cloud data to simulate the uncertainty and data error in practical applications. The 3D origami model is scaled to generate origami patterns of different sizes, thereby improving the robustness of the model to different scales. The 2D origami pattern graph structure is deleted, and edges are added and deleted. Figure 3 As shown in Figure 2, the effect of a 2D origami graph after data augmentation is shown.

[0031] It is important to note that all data augmentation operations remain consistent on the 3D point cloud and the corresponding 2D crease patterns to ensure that the model can capture the true geometric properties of the origami structure.

[0032] Step S2: generating a two-dimensional crease pattern using a graphic diffusion model based on the point cloud data and text description information, and using conditional coding to ensure that the crease pattern conforms to a specified three-dimensional shape; Step S2-1: Using multimodal conditional embedding, the three-dimensional point cloud data and text description are combined to generate conditional embedding data of the graphic diffusion model to control the generation of the crease pattern.

[0033] like Figure 4 As shown, the 3D point cloud features are extracted through the PointNet network, and the text description is encoded using the pre-trained CLIP model to generate semantic embeddings to obtain a high-dimensional feature representation of the point cloud. ,in, is the number of points, is the feature dimension. The text description is encoded by the pre-trained CLIP model to obtain the semantic embedding vector In order to achieve effective fusion between different modalities (point cloud and text), the Cross-Attention Mechanism is introduced. Specifically, the text is embedded As the query vector (Query), point cloud features As key and value vectors. The calculation process of cross-attention is as follows:

[0034] in, is the query vector, is the key vector, is a value vector, 、 、 is the learnable linear transformation parameter matrix, T is the transpose, dk is the scaling factor of the attention score. Through this cross-attention mechanism, the text condition acts as a control signal to guide the weighted aggregation of point cloud features, thereby generating conditional embedding data that integrates semantics. :

[0035] The fusion feature These embeddings are mapped into a unified latent space and used as conditional input to guide the subsequent diffusion model to generate fold patterns that meet semantic constraints. These embeddings are fused through a cross-attention mechanism to map the conditional information into the same latent space, ensuring that the fold patterns generated by the diffusion process meet the specified conditions.

[0036] Step S2-2, embedding random noise and conditions into the input graphic diffusion model; The 3D point cloud data and the corresponding text description are encoded and mapped into the same latent space using the cross-attention mechanism architecture in CLIP. The diffusion process begins with random noise and gradually back-diffusion generates a 2D crease pattern while satisfying the geometric constraints of origami, including fold types such as boundaries, mountain folds, and valley folds. Through the graph diffusion model, a 2D crease pattern that conforms to the desired geometry can be generated.

[0037] Multimodal conditional embedding is used to combine 3D origami point clouds and natural language descriptions, and multimodal conditions are used to control the generation process to ensure that the generated crease pattern is consistent with the input 3D geometric form and meets specific origami parameter requirements.

[0038] Step S2-3: Using a graph-based diffusion model to generate a two-dimensional crease pattern based on the initial random noise.

[0039] In this embodiment, Figure 5 As shown in the figure, a graph-based conditional diffusion model is proposed, which uses the GATv2 (GraphAttention Network v2) graph attention network, pooling layers, and convolutional layers with a stride of 2. The graph-based conditional diffusion model generates a qualified two-dimensional crease pattern from random noise through a graph diffusion mechanism. The specific model structure is as follows: Each graph convolutional layer uses GATv2 to capture the local geometric relationships and topological structure of nodes in the fold pattern. GATv2 adaptively adjusts the weights of neighboring node features, allowing the model to dynamically focus on important geometric relationships. The core mechanism of GATv2 is to assign different weights to nodes through the attention mechanism. The specific calculation formula is as follows:

[0040] in, For nodes In the Characteristics of the layer; For nodes The set of neighbor nodes of is a trainable weight matrix; is the activation function; For nodes and nodes The attention weights between are calculated by the attention mechanism of GATv2. For node j in l Through multi-layer GATv2 graph convolution, the model can capture the multi-scale features of the crease pattern and improve the geometric accuracy of the generated crease pattern.

[0041] The pooling layer is used to reduce the dimensionality of the output features of the graph convolution block, reducing computational effort while preserving key geometric information. Using adaptive pooling, this layer reduces the graph convolution features to a fixed size to ensure consistent feature dimensions and facilitate subsequent convolution operations. Adaptive pooling aggregates node features to extract global information from the graph, making the model more efficient when handling complex geometric structures.

[0042] After GATv2 graph convolution and pooling, the features are fed into a stride-2 convolution layer to further extract global features and improve the model's computational efficiency. The convolution kernel size is 3×3 to capture detailed features of the graph. This convolution operation further reduces the dimensionality of the generated crease pattern features and fuses them, providing support for the final crease pattern generation.

[0043] Specifically, the diffusion model generates a 2D crease pattern starting from random noise. At each diffusion step, the generation process is adjusted through conditional encoding to ensure that the pattern gradually conforms to the 3D target shape and geometric constraints. The inverse formula of the diffusion process is as follows:

[0044] in, Generate results for step t-1, For the Step generation results, is the noise prediction network, and is the noise parameter of the diffusion process, is Gaussian noise, Conditional embeddings composed of text and point cloud data.

[0045] Step S3: Optimize the two-dimensional crease pattern using a physical information neural network to obtain a final two-dimensional origami pattern.

[0046] In this embodiment, Figure 6 As shown in the figure, a physical information neural network (PINN) is used to geometrically and physically optimize the initial two-dimensional crease pattern generated by the diffusion model. By incorporating physical and geometric loss terms, PINN ensures that the generated crease pattern not only meets the geometric structure requirements of actual origami but also can be physically folded. The specific design is as follows: In step S3-1, the input of PINN is the two-dimensional crease pattern output by the diffusion model, including vertex positions, edge (crease) connection information, and the initially generated folding angle.

[0047] In step S3-2, PINN uses a three-layer perceptron (MLP) and a graph convolutional neural network (such as GATv2) to jointly encode nodes and edges, extracting the structural features of the crease pattern. The MLP is used to extract local attribute features, while the GATv2 is used to model the topological relationships between nodes, forming high-level node embeddings and edge features.

[0048] Step S3-3: Input the encoded features into the PINN's physical optimization module. This module incorporates geometric constraints (such as angle closure and edge length consistency) and physical constraints (such as foldability) into its loss function. Through end-to-end optimization, the final output crease pattern is not only geometrically reasonable but also conforms to the physical laws of origami.

[0049] Physical constraints ensure that the crease pattern geometrically meets the basic requirements of origami, such as edge connectivity and fold angles at vertices. The core of this module is the closure constraint, which ensures that the sum of the crease angles formed by each internal vertex in the pattern is close to 2π (360 degrees), thus ensuring that the crease pattern is geometrically closed. The specific loss calculation formula is:

[0050] in, is the set of all internal vertices, is the angle formed by adjacent edges. This loss term ensures that the generated pattern meets the geometric requirements by penalizing the deviation from geometric closure.

[0051] Geometric constraints ensure that the crease pattern can be physically folded. They mainly include the following two constraints: Maekawa constraint and Kawasaki constraint. Loss function for:

[0052] Among them, the Maekawa constraint ensures that the difference between the number of mountain folds and valley folds at each internal vertex is 2, which is one of the necessary conditions for the origami pattern to be physically foldable. The loss term is defined as follows:

[0053] here, and They represent the number of mountain folds and valley folds of vertex i respectively.

[0054] The Kawasaki constraint is used to ensure that the pattern satisfies the Kawasaki theorem, which states that the sum of the relative fold angles at each internal vertex should be equal. The constraints are calculated as follows:

[0055] Where n is the number of creases at the vertex, and represents the relative crease angle. Furthermore, the loss function of the origami design method based on the diffusion model and physical information neural network includes diffusion loss, geometric loss, and physical constraint loss. Diffusion loss controls the reduction of noise during the generation process; geometric loss is used to constrain the vertex positions and edge types of the crease pattern; and physical constraint loss ensures that the generated crease pattern is physically foldable. The total loss function is expressed as:

[0056] in, is the diffusion loss, is a geometric constraint, For physical constraints, and are constants and represent the weights of the loss function.

[0057] in, The specific expression is:

[0058] Among them, E is the mathematical expectation, is the crease structure without diffusion, is the real Gaussian noise, is the noise predicted by the model, is the conditional feature fused from text and point cloud, For the The result is generated by step diffusion.

[0059] In this example, the loss function combines diffusion loss, geometric loss, and physical constraint loss to guide the model in generating crease patterns that both conform to origami geometry and meet physical folding requirements. These three networks form a unified architecture that can be trained in an end-to-end learning approach.

[0060] In this embodiment, model training is performed based on a diffusion model and physical information neural network (PINN) origami design method, configuring the data path, pre-trained model path, model storage path, and initialization parameters (such as learning rate, regularization term, and optimization algorithm). During training, a 3D origami model and its corresponding conditional description (including point cloud and text information) are randomly extracted from the training dataset and input into the multimodal conditional encoding module. The diffusion model receives initial random noise and, guided by multimodal conditions, gradually generates a preliminary 2D crease pattern. The graph convolution diffusion model uses the GATv2 graph convolutional network to extract multi-scale graph features and generates the crease pattern through a multi-step back-diffusion process. The generated preliminary crease pattern is input into the physical information neural network (PINN), where it is optimized using geometric and physical constraints to ensure that the pattern conforms to the physical and geometric rules of origami. The entire model continuously optimizes parameters through back-propagation, ultimately outputting a physically feasible and geometrically accurate 2D origami pattern.

[0061] In this embodiment, model testing is similar to the training process, using 3D point cloud data and conditional descriptions. During the testing phase, parameters such as the model path, test data path, and output path are set. A preliminary crease pattern is generated using the diffusion model and fed into the PINN for optimization, ultimately outputting a 2D origami pattern that meets the requirements. The generated 2D crease pattern is verified for its physical feasibility and geometric accuracy through simulated folding processes.

[0062] The diffusion model uses random noise as input and generates preliminary two-dimensional crease patterns through a multi-step back-diffusion process. A physically-informed neural network (PINN) receives these preliminary patterns and optimizes them using geometric and physical constraints.

[0063] The model uses a multimodal conditional encoding (combining 3D point clouds and text descriptions) to guide the diffusion process. During each training session, a 3D origami model and its conditional description are randomly sampled from the dataset. The model then optimizes the generated pattern through backpropagation, gradually making it conform to the target geometry and physics requirements.

[0064] The physically-informed neural network optimizes the preliminary crease pattern using geometric and physical loss terms to ensure that the pattern is physically foldable and geometrically accurate. The loss function incorporates geometric closure, the Maekawa and Kawasaki theorems, and other factors into the optimization process.

[0065] During training, in one embodiment, the diffusion model and PINN are jointly trained and optimized, with model parameters updated at each training round to ensure that the generated crease patterns gradually meet origami design requirements. Training is performed using the AdamW optimizer, with a learning rate decay strategy to accelerate convergence.

[0066] In this embodiment, Figure 7 Figure 2 shows the generated 2D origami pattern and its performance in actual folding. The test results include the geometric accuracy of the crease pattern, its physical feasibility, and the degree of consistency between the actual folded 3D shape and the original input shape.

[0067] Step S4: File Storage and Visual Display. The generated two-dimensional origami pattern is stored in the computer's memory, and a visualization tool is used to display the crease pattern generation process and its physical folding effects. The user can select different visualization modes to more intuitively understand the folding process and results of the generated pattern.

[0068] This origami inverse design system based on diffusion model and physical information neural network provides an efficient and reliable method to generate complex two-dimensional crease patterns, which is widely applicable to fields such as materials science, robotics and biomedical engineering.

[0069] like Figure 8 As shown, the technical solution of this embodiment ensures the accuracy of the origami inverse process design, where the blue vertices are the original origami model graph structure and the yellow graph is the origami pattern graph structure inferred by multiple sampling.

[0070] Example 2 In one or more embodiments, an origami design system based on a diffusion model and a physical information neural network is disclosed, specifically comprising: a data acquisition module configured to: acquire point cloud data of a target three-dimensional origami shape and corresponding text description information; a preliminary crease generation module configured to: generate a two-dimensional crease pattern using a graphic diffusion model based on the point cloud data and text description information, and ensure that the crease pattern conforms to a specified three-dimensional shape using conditional coding; The origami generation module is configured to optimize the two-dimensional crease pattern using a physical information neural network to obtain a final two-dimensional origami diagram.

[0071] Example 3 This embodiment provides an electronic device, including a memory and a processor, and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, the steps of the origami design method based on the diffusion model and the physical information neural network are completed.

[0072] Example 4 This embodiment provides a computer-readable storage medium for storing computer instructions. When the computer instructions are executed by a processor, the steps of the origami design method based on the diffusion model and the physical information neural network are completed.

[0073] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts 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, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0074] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0075] These computer program instructions can also be loaded onto a computer or other programmable data processing device, and a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide the functions for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0076] The descriptions of the various embodiments in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0077] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. The origami design method based on diffusion model and physical information neural network is characterized by: include: Obtain point cloud data and corresponding text description information of the target three-dimensional origami shape; Generate a two-dimensional crease pattern using a graphic diffusion model based on the point cloud data and text description information, and use conditional coding to ensure that the crease pattern conforms to a specified three-dimensional shape; The two-dimensional crease pattern is optimized using a physical information neural network to obtain a final two-dimensional origami figure.

2. The origami design method based on the diffusion model and physical information neural network according to claim 1, characterized in that: The conditional encoding uses multimodal conditional embedding to combine 3D point cloud data and text description to generate conditional embedding data for the graph diffusion model. Specifically: The 3D point cloud features are extracted through the PointNet network, and the text description is encoded using the pre-trained CLIP model to generate semantic embeddings; the extracted 3D point cloud features and semantic embeddings are fused through the cross-attention mechanism, mapping the conditional information to the same latent space to obtain conditional embedding data.

3. The origami design method based on diffusion model and physical information neural network according to claim 1, characterized in that: The input of the graph diffusion model is random noise and conditional embedding data.

4. The origami design method based on diffusion model and physical information neural network according to claim 1, characterized in that: The graph diffusion model includes a graph attention network, a pooling layer, and a convolutional layer with a stride of 2; The graph attention network assigns different weights to nodes through the attention mechanism, which is expressed as: in, For nodes In the Characteristics of the layer; For nodes The set of neighbor nodes of is a trainable weight matrix; is the activation function; For nodes and nodes The attention weights between are calculated by the attention mechanism of GATv2. For node j in l Characteristics of the layer.

5. The origami design method based on diffusion model and physical information neural network according to claim 4, characterized in that: The graphic diffusion model generates a two-dimensional crease pattern starting from random noise, and adjusts the generation process through conditional coding. The inverse formula of the diffusion process is as follows: in, Generate results for step t-1, For the Step generation results, is the noise prediction network, and is the noise parameter of the diffusion process, is Gaussian noise, Conditional embeddings composed of text and point cloud data.

6. The origami design method based on diffusion model and physical information neural network according to claim 1, characterized in that: The physical information neural network includes a three-layer perceptron and a graph convolutional neural network; the physical information neural network uses a loss function to control the reduction of noise during the generation process, constrains the vertex positions and edge types of the crease pattern, and ensures that the generated crease pattern is physically foldable.

7. The origami design method based on diffusion model and physical information neural network according to claim 6, characterized in that: The loss function includes diffusion loss, geometric loss and physical constraint loss; The loss function is expressed as: in, is the diffusion loss, is the geometric loss, is the physical constraint loss, and are constants, representing the weights of the loss function; The specific expression of diffusion loss is: Among them, E is the mathematical expectation, is the crease structure without diffusion, is the real Gaussian noise, is the noise predicted by the model, F cond Conditional features for text and point cloud fusion, For the The result is generated by step diffusion.

8. The origami design system based on diffusion model and physical information neural network is characterized by: include: a data acquisition module configured to: acquire point cloud data of a target three-dimensional origami shape and corresponding text description information; a preliminary crease generation module configured to: generate a two-dimensional crease pattern using a graphic diffusion model based on the point cloud data and text description information, and ensure that the crease pattern conforms to a specified three-dimensional shape using conditional coding; The origami generation module is configured to optimize the two-dimensional crease pattern using a physical information neural network to obtain a final two-dimensional origami diagram.

9. An electronic device, characterized in that: The invention comprises a memory and a processor and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, the origami design method based on the diffusion model and the physical information neural network as described in any one of claims 1 to 7 is completed.

10. A computer-readable storage medium, characterized in that Used to store computer instructions, which, when executed by a processor, complete the origami design method based on the diffusion model and physical information neural network as described in any one of claims 1 to 7.

Images