Object generation design system

Abstract

The present application relates to a generative object design system. A computer-implemented method, when executed by data processing hardware, causes the data processing hardware to perform operations. The operations include providing training data at a diffusion model of a generative object design algorithm, the training data comprising a plurality of stored three-dimensional (3D) objects and a plurality of stored object slices, generating, via the diffusion model, a two-dimensional (2D) object image based on the training data, predicting, via an attention residual U-network model of the generative object design algorithm, a 2D object slice of the generated 2D object image, and combining, via the generative object design algorithm, the predicted object slice into a 3D object design.

Description

[0001] Foreword

[0002] The information provided in this section is for the purpose of presenting the general context of this disclosure. The work of the currently named inventors, to the extent described in this section, and aspects that may not otherwise qualify as prior art at the time of filing, are neither expressly nor implicitly acknowledged as prior art to this disclosure. Technical Field

[0003] This disclosure generally relates to generative three-dimensional (3D) object design systems. Background Technology

[0004] The design and development of vehicle objects (such as wheels) is a complex and time-consuming process that typically involves multiple iterative steps. Designers often spend a significant amount of time refining their concepts through numerous iterations to achieve the desired aesthetic and functional characteristics. This extensive process can limit the number of designs that can be developed within a given timeframe, thus restricting the diversity and innovation of object designs available on the market.

[0005] Once the initial 2D design is complete, it must be converted into a 3D model to assess its feasibility and prepare it for manufacturing. However, current methods for creating these 3D models from 2D sketches or designs are often inefficient and result in outputs lacking the necessary precision. This further lengthens the development timeline and can lead to additional revisions and refinements, increasing the overall effort required to bring new object designs to market. Therefore, an improved method is needed for developing 3D object designs from 2D sketches or images. Summary of the Invention

[0006] In some aspects, the computer-implemented method causes the data processing hardware to perform operations when executed by the data processing hardware. These operations include providing training data at a diffusion model of a generative object design algorithm, the training data comprising multiple stored three-dimensional (3D) objects and multiple stored object slices; generating a two-dimensional (2D) object image via the diffusion model based on the training data; predicting 2D object slices of the generated 2D object image via an attention residual U-network model of the generative object design algorithm; and combining the predicted object slices into a 3D object design via the generative object design algorithm.

[0007] Optionally, the operation may also include generating 3D object components via a generative object design algorithm and extrapolating a 3D object design from the generated 3D object components. In some examples, providing training data may include adjusting a diffusion model based on the training data; the diffusion model is a machine learning model. In some cases, predicting 2D object slices includes generating at least twelve object slices. The operation may also include importing mesh points based on the predicted 2D object slices into a preprocessor of data processing hardware via a generative object design algorithm. Optionally, combining the predicted 2D object slices may include refining the imported mesh points. The operation may further include extrude bottom mesh points along each predicted 2D object slice via a generative object design algorithm. In some cases, the operation may include meshing the 3D object design. The operation may also include performing structural and modal analysis on the meshed 3D object design.

[0008] In other aspects, the generative 3D design system includes data processing hardware and memory hardware that communicates with the data processing hardware. The memory hardware stores instructions that, when executed on the data processing hardware, cause the data processing hardware to perform operations. These operations include providing training data at a diffusion model of the generative object design algorithm, the training data including multiple stored 3D objects and multiple stored object slices; generating 2D object images via the diffusion model based on the training data; predicting object slices from the generated 2D object images via an attention residual U-network model of the generative object design algorithm; and combining the predicted object slices into a 3D object design via the generative object design algorithm.

[0009] In some examples, the stored 3D object may include a rim. Optionally, providing training data may include adjusting a diffusion model based on the training data; the diffusion model is a machine learning model. In some cases, predicting 2D object slices may include generating at least twelve object slices. The operation may also include extrude bottom mesh points along each predicted 2D object slice via a generative object design algorithm. The operation may further include meshing the 3D object design. Optionally, the operation may include performing structural and modal analysis on the meshed 3D object design.

[0010] In a further aspect, the generative 3D design system includes data processing hardware and memory hardware communicating with the data processing hardware. The memory hardware stores instructions that, when executed on the data processing hardware, cause the data processing hardware to perform operations. These operations include providing training data at a diffusion model of a generative object design algorithm, the training data including multiple stored 3D objects and multiple stored object slices; generating a 2D object image via the diffusion model based on the training data; and predicting object slices of the generated 2D object image via an attention residual U-network model of the generative object design algorithm. The operations also include importing mesh points based on the predicted 2D object slices into a preprocessor of the data processing hardware via the generative object design algorithm; combining the predicted object slices into a 3D object design via the generative object design algorithm; meshing the 3D object design; and performing structural and modal analysis on the meshed 3D object design.

[0011] In some examples, the operation may include generating 3D object components via a generative object design algorithm and extrapolating 3D object designs from the generated 3D object components. Attached Figure Description

[0012] The accompanying drawings described herein are for illustrative purposes only for the selected configurations and are not intended to limit the scope of this disclosure.

[0013] Figure 1 This is a schematic diagram of a generative three-dimensional (3D) object design system according to the present disclosure;

[0014] Figure 2 This is an exemplary block diagram of a system for designing generated objects according to this disclosure;

[0015] Figure 3 It is a schematic diagram of the diffusion model of the object generation design system according to the present disclosure, which generates multiple two-dimensional object images;

[0016] Figure 4 This is a schematic diagram of an object slice of a two-dimensional object according to this disclosure;

[0017] Figure 5 This is a schematic diagram of a training database for a system for designing object generation according to this disclosure;

[0018] Figure 6 This is a schematic diagram illustrating a generative object design system that uses object slicing to generate 3D objects;

[0019] Figure 7 This is a schematic diagram of multiple grid points and grid functions of the object generation design system based on this disclosure;

[0020] Figure 8 It is a meshed 3D object according to this disclosure;

[0021] Figure 9 This is an example of a diffusion model for generating two-dimensional objects according to this disclosure;

[0022] Figure 10 This is an example of an attention residual U-network model for generating multiple object slices according to this disclosure;

[0023] Figure 11 This is an example of a diffusion model for generating a portion of a two-dimensional object according to this disclosure;

[0024] Figure 12 This is an example of an attention residual U-network model for generating multiple object slices according to this disclosure;

[0025] Figure 13 Generative object design algorithms utilize Figure 11 Examples of generating 3D objects by slicing multiple objects; and

[0026] Figure 14 This is an exemplary method of operation for designing an object generation system based on this disclosure.

[0027] Throughout the accompanying drawings, corresponding reference numerals indicate the corresponding parts. Detailed Implementation

[0028] The example configuration will now be described more fully with reference to the accompanying drawings. The example configuration is provided so that this disclosure will be thorough and will fully communicate the scope of this disclosure to those skilled in the art. Specific details, such as examples of specific components, devices, and methods, are set forth to provide a thorough understanding of the configuration of this disclosure. It will be apparent to those skilled in the art that the specific details are not required, the example configuration can be implemented in many different forms, and the specific details and example configuration should not be construed as limiting the scope of this disclosure.

[0029] The terminology used herein is for the purpose of describing a particular exemplary configuration only and is not intended to be restrictive. As used herein, the singular articles “a,” “an,” and “the” may be intended to include plural forms as well, unless the context clearly indicates otherwise. The terms “comprises,” “comprising,” “including,” and “having” are inclusive and therefore specify the presence of features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof. Unless specifically identified as an order of execution, the method steps, processes, and operations described herein should not be construed as requiring them to be performed in the specific order discussed or shown. Additional or alternative steps may be employed.

[0030] When an element or layer is referred to as being “on,” “joined to,” “connected to,” “attached to,” or “linked to” another element or layer, it may be directly on, joined to, connected to, attached to, or linked to the other element or layer, or there may be intermediate elements or layers present. Conversely, when an element is referred to as being “directly on,” “directly joined to,” “directly connected to,” “directly attached to,” or “directly linked to” another element or layer, there may be no intermediate elements or layers present. Other terms used to describe relationships between elements should be interpreted in a similar manner (e.g., “between” vs. “directly between,” “adjacent” vs. “directly adjacent,” etc.). As used herein, the term “and / or” includes any and all combinations of one or more of the associated listed items.

[0031] The terms “first,” “second,” “third,” etc., may be used herein to describe various elements, components, regions, layers, and / or parts. These elements, components, regions, layers, and / or parts should not be limited by these terms. These terms may be used only to distinguish one element, component, region, layer, or part from another. Unless the context clearly indicates otherwise, terms such as “first,” “second,” and other numerical terms do not imply order or sequence. Therefore, without departing from the teachings of the example configuration, the first element, component, region, layer, or part discussed below may be referred to as the second element, component, region, layer, or part.

[0032] In this application, including the following definitions, the term "module" may be replaced by the term "circuit". The term "module" may refer to, be a part of, or include: application-specific integrated circuits (ASICs); digital, analog, or mixed-signal analog / digital discrete circuits; digital, analog, or mixed-signal analog / digital integrated circuits; combinational logic circuits; field-programmable gate arrays (FPGAs); processors (shared, dedicated, or grouped) that execute code; memories (shared, dedicated, or grouped) that store code executed by the processor; other suitable hardware components that provide the described functionality; or combinations of some or all of the foregoing, such as in a system-on-a-chip.

[0033] The term "code" as used above can include software, firmware, and / or microcode, and can refer to programs, routines, functions, classes, and / or objects. The term "shared processor" covers a single processor that executes some or all of the code from multiple modules. The term "group processor" covers a processor that, in combination with additional processors, executes some or all of the code from one or more modules. The term "shared memory" covers a single memory that stores some or all of the code from multiple modules. The term "group memory" covers memory that, in combination with additional memory, stores some or all of the code from one or more modules. The term "memory" can be a subset of the term "computer-readable medium." The term "computer-readable medium" does not include transient electrical and electromagnetic signals propagating through the medium, and therefore can be considered tangible and non-transitory memory. Non-limiting examples of non-transitory memory include tangible computer-readable media, which include non-volatile memory, magnetic memory, and optical memory.

[0034] The apparatus and methods described in this application can be implemented, in part or in whole, by one or more computer programs executed by one or more processors. The computer program includes processor-executable instructions stored on at least one non-transitory tangible computer-readable medium. The computer program may also include and / or depend on stored data.

[0035] A software application (i.e., a software resource) can refer to computer software that enables a computing device to perform tasks. In some examples, a software application may be referred to as an "application," "app," or "program." Example applications include, but are not limited to, system diagnostic applications, system management applications, system maintenance applications, word processing applications, spreadsheet applications, messaging applications, media streaming applications, social networking applications, and game applications.

[0036] Non-transitory memory can be a physical device used to temporarily or permanently store programs (e.g., instruction sequences) or data (e.g., program state information) for use by a computing device. Non-transitory memory can be volatile and / or non-volatile addressable semiconductor memory. Examples of non-volatile memory include, but are not limited to, flash memory and read-only memory (ROM) / programmable read-only memory (PROM) / erasable programmable read-only memory (EPROM) / electrically erasable programmable read-only memory (EEPROM) (e.g., commonly used in firmware, such as bootloaders). Examples of volatile memory include, but are not limited to, random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), phase-change memory (PCM), and magnetic disks or magnetic tapes.

[0037] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for a programmable processor and can be implemented using high-level procedural and / or object-oriented programming languages ​​and / or assembly / machine languages. As used herein, the terms “machine-readable medium” and “computer-readable medium” refer to any computer program product, non-transitory computer-readable medium, apparatus, and / or device (e.g., disk, optical disk, memory, programmable logic device (PLD)) used to provide machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term “machine-readable signal” refers to any signal used to provide machine instructions and / or data to a programmable processor.

[0038] Various implementations of the systems and techniques described herein can be implemented in digital electronic and / or optical circuits, integrated circuits, specially designed ASICs (Application-Specific Integrated Circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementations in one or more computer programs executable and / or interpretable on a programmable system, which includes at least one programmable processor, which may be dedicated or general-purpose, coupled to receive data and instructions from a storage system, at least one input device, and at least one output device, and to transmit data and instructions to the storage system, at least one input device, and at least one output device.

[0039] The processes and logic described in this specification can be executed by one or more programmable processors (also known as data processing hardware) that execute one or more computer programs to perform functions by manipulating input data and generating output. The processes and logic can also be executed by special-purpose logic circuitry, such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits). Processors suitable for executing computer programs include, for example, both general-purpose microprocessors and special-purpose microprocessors, as well as any one or more processors of any kind of digital computer. Typically, the processor receives instructions and data from read-only memory or random access memory, or both. The basic elements of a computer are a processor for executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices (e.g., magnetic disks, magneto-optical disks, or optical disks) for storing data, or operatively coupled to receive data from or transfer data to one or more mass storage devices, or both. However, a computer does not need to have such devices. Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, including, for example, semiconductor memory devices such as EPROM, EEPROM, and flash memory devices; magnetic disks, such as internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. Processors and memory may be supplemented by or incorporated into dedicated logic circuitry.

[0040] To provide interaction with a user, one or more aspects of this disclosure can be implemented on a computer having a display device for displaying information to the user, such as a CRT (cathode ray tube), LCD (liquid crystal display) monitor, or touchscreen, and optionally a keyboard and pointing device, such as a mouse or trackball, through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback; and input from the user can be received in any form, including acoustic, voice, or tactile input. Furthermore, the computer can interact with the user by sending documents to and receiving documents from the device used by the user; for example, by sending a webpage to a web browser on the user's client device in response to a request received from a web browser.

[0041] refer to Figures 1 to 5The generative object 3D design system 10 is a computer-implemented system including a controller 12 that executes the generative object design algorithm 14. For example, the generative object design system 10 can be configured as a web-based application with an interactive user interface 100. Users can utilize the interactive user interface 100 to import or otherwise upload images 102. For example, users can drag and drop or manually upload images 102 via the interactive user interface 100. The generative object design system 10 automatically resizes the images 102 to a standard format for consistent processing and prediction, as described herein.

[0042] The object design algorithm 14 is executed by the data processing hardware 16 of the controller 12. The controller 12 may be a computer or other processor configured with the data processing hardware 16. The controller 12 also includes memory hardware 18 that communicates with the data processing hardware 16. The memory hardware 18 stores instructions that, when executed by the data processing hardware 16, cause the data processing hardware 16 to perform the operations described herein. The object design algorithm 14 is configured to generate a three-dimensional (3D) object model 20 based on predicted two-dimensional (2D) object slices 22, which will be described in more detail below. The user may store the predicted 2D object slices 22 in the memory hardware 18 and / or may use the predicted 2D object slices 22 to generate the 3D object model 20.

[0043] Further reference Figure 1-5 The object design algorithm 14 includes a diffusion model 30 configured as a machine learning model. The object design algorithm 14 trains the diffusion model 30 using training data 32. The diffusion model 30 includes a model trainer 34 configured to acquire or receive training data 32 for training the diffusion model 30. As described above, the diffusion model 30 is configured as a machine learning model 30 such that the model trainer 34 is configured to train the machine learning model 30 based on the training data 32. The model trainer 34 can receive the training data 32 from, for example, a training database 36 stored in memory hardware 18. The training data 32 includes, but is not limited to, multiple stored 3D objects 32a and stored object slices 32b. The training data 32 may also include any type of data (text, images, audio, etc.) that the diffusion model 30 is trained to receive. For example, the training data 32 may include text input 32c received by the diffusion model 30. The diffusion model 30 is trained to predict and generate 2D object images 38, which can be used to ultimately generate the 3D object model 20.

[0044] During the training of the diffusion model 30, training data 32 is provided to the diffusion model 30. As described above, the training data 32 may include stored 3D objects 32a. For example, the stored 3D objects 32a may include dozens, hundreds, and / or thousands of object images stored in the memory hardware 18. The stored 3D objects 32a may correspond to, but are not limited to, rims. For example, a rim may be an easily accessible image that can be uploaded as training data 32 and provided to the diffusion model 30. Stored object slices 32b are slices of the stored 3D objects 32a. For example, stored object slices 32b may include multiple slices of the rims of the stored 3D objects 32a. Each stored 3D object 32a may have a corresponding stored object slice 32b. As a result, the training database 36 has a wide array of variations in the stored 3D objects 32a and stored object slices 32b that can be used to train the diffusion model 30.

[0045] The stored object slice 32b can typically contain approximately twelve (12) slices of the corresponding stored 3D object image 32a. The number of slices in the stored object slice 32b can vary depending on the complexity of the stored 3D object 32a. Using approximately twelve (12) slices allows for advantageous training of the diffusion model 30 on a wide array of object segments, which is advantageous for generating the 3D object model 20. As described in more detail below, the diffusion model 30 can be used at least partially to assist in generating the 3D object model 20, in that the diffusion model 30 can be used to generate one or more 2D object images 38. For example, Figure 3 A diffusion model 30 is shown, which receives a text prompt 32c and, in response, generates multiple 2D object images 38. The text prompt 32c can provide context to the diffusion model 30, allowing the diffusion model 30 to use the context from the text prompt 32c to parse training data 32 and generate 2D object images 38 based on the training data 32. Although the diffusion model 30 uses training data 32 to inform the generated 2D object images 38, the 2D object images 38 are a new design independent of the training data 32.

[0046] For example, the diffusion model 30 uses training data 32 to inform various 2D object images 38, in that the diffusion model 30 can extract aspects from different sets of training data 32 (i.e., stored 3D objects 32a and stored object slices 32b). The diffusion model 30 can further utilize text prompts 32c in conjunction with the aspects extracted from the training data 32 to generate the 2D object image 38. The text prompts 32c may indicate user preferences or other design contexts that can be requested by the user. In some cases, the text prompts 32c may include an image 102 that can be input at the interactive user interface 100. The diffusion model 30 can utilize the image 102 to select various combinations of stored 3D object models 32a and stored object slices 32b to generate the 2D object image 38.

[0047] refer to Figure 2-10 A 2D object image 38 is provided to an attention residual U-network model 40 of the object generation design algorithm 14 to predict object slices 22 of the generated 2D object image 38. The attention residual U-network model 40 receives one or more 2D object images 38 and is configured to predict 2D object slices 22 for each 2D object image 38. For example, a user can select one of the 2D object images 38 generated by the diffusion model 30. Based on the user selection, the attention residual U-network model 40 predicts and generates 2D object slices 22. The attention residual U-network model 40 collaborates with the diffusion model 30 as part of the object generation design algorithm 14 to predict 2D object slices 22. For example, the attention residual U-network model 40 may have similar access to a training database 36 to compare the generated 2D object image 38 with stored 3D object models 32a and stored image slices 32b. Based on this comparison, the attention residual U-network model 40 can predict individual object slices 22 of the selected 2D object image 38.

[0048] In some cases, the attention residual U-network model 40 can receive an image 102 input at the interactive user interface 100 and can use the image 102 to inform the predicted 2D object slices 22. The predicted 2D object slices 22 can be presented as a series of black and white and / or grayscale images illustrating the 2D object images 38 generated at each slice 22. The object design generation algorithm 14 can use the black and white predicted 2D object slices 22 to generate 3D object designs, as the contrast between black and white can provide improved edge detection. As mentioned above, the predicted 2D object slices 22 can be stored in memory hardware 18 and / or can be stored locally by the user.

[0049] Further reference Figure 2-10The object design algorithm 14 extracts pixel coordinates 50 from the predicted 2D object slices 22. The pixel coordinates 50 correspond to objects in the predicted 2D object slices 22, allowing the object design algorithm 14 to ignore pixels that may not correspond to objects in the predicted 2D object slices 22. The object design algorithm 14 uses the pixel coordinates 50 to create corresponding grid points 52 for each valid pixel coordinate 50. The object design algorithm 14 can also apply a scaling factor 54 to each of the 2D object slices 22 to ensure that the 2D object slices 22 accurately reconstruct objects when stacked. The object design algorithm 14 communicates the grid points 52 with the preprocessor 60 of the data processing hardware 16 for further processing.

[0050] Grid point 52 captures the features of the object. For example, if the object is a wheel, the features may include, but are not limited to, the rim, spokes, center hole area, lug nut hole, and outer lip. Figure 8 An example of a 3D object model 20 (i.e., a wheel) is shown after the preprocessor 60 has refined the mesh points 52. Figure 7 The refinement of mesh point 52 is shown. For example, mesh points 52 from each slice 22 are wrapped with solid hexahedral elements to form a mesh 56 for each slice 22. The bottom mesh point 52 of each slice 22 is along the object centerline C. 20 Extrusion is used to fill the gaps 58 defined between each slice 22. For example, mesh 56 fills the gaps 58 to provide improved resolution of object model 20 while capturing details of object model 20.

[0051] Therefore, as a result of the refined mesh points 52 and the meshed object model 20 from the preprocessor 60, the predicted 2D object slices 22 are combined via the generative object design algorithm 14. The combined predicted 2D object slices 22 produce the 3D object model 20. Once the 3D object model 20 is meshed and generated, the generative object design algorithm 14 can perform a smoothing function 70, which is configured to eliminate roughness and steps from the process construction of the 3D object model 20 from the predicted 2D object slices 22. Once smoothed, the generative object design algorithm 14 performs structural and modal analysis 72 on the 3D object design 20. The structural and modal analysis 72 is used to apply loads and constraints to the 3D object model 20 to evaluate the usability of the 3D object model 20.

[0052] A diffusion model 30 is used to generate a 2D object image 38. For example, Figure 9 The output of the 2D object image 38 after executing the diffusion model 30 is shown. The 2D object image 38 can then be fed into the attention residual U-network model 40 to output predicted 2D object slices 22. The 2D object slices 22 are projected slices based on the generated 2D object image 38. For example, Figure 10 The output of 2D object slice 22 generated by attention residual U-net model 40 is shown.

[0053] For details, please refer to the following: Figure 2 and 11 -13, the object design algorithm 14 can be used to generate individual 3D object components 20a of the 3D object model 20. For example, Figures 11 to 13 The diagram illustrates the use of generative object design algorithm 14 to generate a single spoke 20a of wheel model 20. Generative object design algorithm 14 utilizes a process similar to that described above. Generative object design algorithm 14 can execute diffusion model 30 to generate a 2D object image 38a. Diffusion model 30 still utilizes training database 36 to train and generate the 2D object image 38a, regardless of whether diffusion model 30 is used to ultimately generate a single 3D object component 20a or the entire 3D object model 20.

[0054] Once the 2D object image 38a is generated, the object generation design algorithm 14 executes the attention residual U-network model 40. The attention residual U-network model 40 predicts and generates 2D object slices 22a based on the generated 2D object image 38a from the diffusion model 30. In the example shown ( Figure 11 In the process, 2D object slices 22a capture the various components 20a of object model 20. The object generation design algorithm 14 combines the 2D object slices 22a into 3D object components 20a. The object generation design algorithm 14 extracts pixel coordinates 50 from slices 22a and creates corresponding mesh points 52. The mesh points 52 are then... Figure 2-10 The described process communicates with the preprocessor 60. The object design algorithm 14 can then form a mesh 56 for each of the slices 22a based on the mesh points 52 to generate 3D object components 20a.

[0055] The object design algorithm 14 can then perform a smoothing function 70 and run structural and modal analysis 72 on the 3D object component 20a. The object design algorithm 14 can also utilize the 3D object component 20a to infer and form the entirety of the 3D object model 20. For example, Figure 12 An object design algorithm 14 for generating a 3D object model 20 from a 3D object component 20a is shown.

[0056] refer to Figure 14An example method 1400 of the operation of the generative object design system 10 is shown. At 1402, training data 32 is provided to the diffusion model 30 of the generative object design algorithm 14. The training data 32 includes multiple stored three-dimensional (3D) objects 32a and multiple stored object slices 32b. At 1404, the diffusion model 30 generates a two-dimensional (2D) object image 38 based on the training data 32. At 1406, the attention residual U network model 40 of the generative object design algorithm 14 predicts object slices 22 of the generated 2D object image 38. At 1408, based on the predicted 2D object slices 22, grid points 52 are imported into the preprocessor 60 of the data processing hardware 16 via the generative object design algorithm 14. At 1410, the generative object design algorithm 14 combines the predicted object slices 22 into a 3D object model 20. The object design algorithm 14 meshes the 3D object model 20 at 1412 and performs structural and modal analysis 72 on the meshed 3D object model 20 at 1414.

[0057] Many embodiments have been described. However, it should be understood that various modifications can be made without departing from the spirit and scope of this disclosure. Therefore, other embodiments are within the scope of the appended claims.

[0058] The foregoing description has been provided for purposes of illustration and description. It is not intended to be exhaustive or limiting of this disclosure. Elements or features of a particular configuration are generally not limited to that particular configuration, but are interchangeable where applicable and can be used in selected configurations, even if not specifically shown or described. They can also be varied in many ways. Such variations should not be considered as departing from this disclosure, and all such modifications are intended to be included within the scope of this disclosure.

Claims

1. A computer-implemented method, when executed by data processing hardware, causes the data processing hardware to perform an operation, the operation comprising: Training data is provided at the diffusion model of the generative object design algorithm, the training data including multiple stored three-dimensional (3D) objects and multiple stored object slices; Based on the training data, a two-dimensional (2D) object image is generated via the diffusion model; The attention residual U-network model of the generated object design algorithm is used to predict 2D object slices of the generated 2D object image; and The predicted object slices are combined into a 3D object design using the generated object design algorithm.

2. The method according to claim 1, further comprising generating 3D object components via the generated object design algorithm, and extrapolating the 3D object design from the generated 3D object components.

3. The method of claim 1, wherein providing the training data includes adjusting the diffusion model based on the training data, the diffusion model being a machine learning model.

4. The method of claim 1, wherein predicting the 2D object slices comprises generating at least twelve object slices.

5. The method of claim 1, further comprising importing mesh points based on predicted 2D object slices into a preprocessor of the data processing hardware via the generated object design algorithm.

6. The method of claim 5, wherein the combined predicted 2D object slices include refined imported mesh points.

7. The method of claim 6, further comprising extruded bottom grid points along each predicted 2D object slice via a generative object design algorithm.

8. The method of claim 7, further comprising meshing the 3D object design.

9. The method according to claim 8 further includes performing structural and modal analysis on the meshed 3D object design.

10. A design system for generating three-dimensional (3D) objects, configured to perform the method of claim 1.

Images