Autodesk, Inc.

Within Technologies Ltd.

Grow Software Limited

Hooman Shayani

Konara Mudiyanselage Kosala Bandara

Aditya Sanghi

Mark Thomas Davis

Adrian Adam Thomas Butscher

Andriy Banadyha

Joseph George Lambourne

Marco Fumero

Karl Darcy Daniel Willis

Andreas Linas Bastian

Andrew John Harris

Nigel Jed Wesley Morris

Andriy Banadyga

Nathan David Rogers

Siavash Haroun Mahdavi

Stephen Barley

Pradeep Kumar Jayaraman

Martin Raymond Razzell

Luke Edwards

Daniele Grandi

Nandakumar Santhanam

Thomas Davies

Martin Cvetanov Marinov

Xuejun Sheng

Jingyang John Chen

Anthony Ruto

Marco Amagliani

Suguru Furuta

Anthony Christopher Kipkirui Yegon Ruto

Dagmara Lilianna Szkurlat

Brian M Frank

Michael Andrew Smell

Tristan Ward Barback

Robert Bolter

Christopher Michael Wade

Peter Meltzer

Amir Hosein Khas Ahmadi

Srikanth Reddy Bethi

In various embodiments, a style comparison metric application generates a style comparison metric for pairs of different three dimensional (3D) computer-aided design (CAD) objects. In operation, the style comparison metric application executes a trained neural network any number of times to map 3D CAD objects to feature maps. Based on the feature maps, the style comparison metric application computes style signals. The style comparison metric application determines values for weights based on the style signals. The style comparison metric application generates the style comparison metric based on the weights and a parameterized style comparison metric.

Techniques for generating subjective style comparison metrics for B-reps of 3D CAD objects

In various embodiments, a parameter domain graph application generates UV-net representations of 3D CAD objects for machine learning models. In operation, the parameter domain graph application generates a graph based on a B-rep of a 3D CAD object. The parameter domain graph application discretizes a parameter domain of a parametric surface associated with the B-rep into a 2D grid. The parameter domain graph application computes at least one feature at a grid point included in the 2D grid based on the parametric surface to generate a 2D UV-grid. Based on the graph and the 2D UV-grid, the parameter domain graph application generates a UV-net representation of the 3D CAD object. Advantageously, generating UV-net representations of 3D CAD objects that are represented using B-reps enables the 3D CAD objects to be processed efficiently using neural networks.

Techniques for generating UV-net representations of 3D CAD objects for machine learning models

One embodiment of the present invention sets forth a technique for performing style transfer. The technique includes determining a distribution associated with a plurality of style codes for a plurality of three-dimensional (3D) shapes, where each style code included in the plurality of style codes represents a difference between a first 3D shape and a second 3D shape, and where the second 3D shape is generated by applying one or more augmentations to the first 3D shape. The technique also includes sampling from the distribution to generate an additional style code and executing a trained machine learning model based on the additional style code to generate an output 3D shape having style-based attributes associated with the additional style code and content-based attributes associated with an object. The technique further includes generating a 3D model of the object based on the output 3D shape.

Generating styles for neural style transfer in three-dimensional shapes

Methods, systems, and apparatus, including medium-encoded computer program products, for computer aided design of physical structures using generative design processes. A method includes obtaining one or more design criteria for a modeled object including at least one design constraint; calculating a series of target values for the at least one design constraint, from an initial target value to a final target value; iteratively modifying a generatively designed three dimensional shape of the modeled object in the design space, wherein the iteratively modifying comprises performing numerical simulation of the modeled object, computing shape change velocities for an implicit surface in a level-set representation of the three dimensional shape in accordance with respective ones of target values in the series of target values, starting from the initial target value and ending with the final target value, and updating the level-set representation using the shape change velocities.

Generative design shape optimization with controlled convergence for computer aided design and manufacturing

One embodiment of the present invention sets forth a technique for performing style transfer. The technique includes generating an input shape representation that includes a plurality of points near a surface of an input three-dimensional (3D) shape, where the input 3D shape includes content-based attributes associated with an object. The technique also includes determining a style code based on a difference between a first latent representation of a first 3D shape and a second latent representation of a second 3D shape, where the second 3D shape is generated by applying one or more augmentations to the first 3D shape. The technique further includes generating, based on the input shape representation and style code, an output 3D shape having the content-based attributes of the input 3D shape and style-based attributes associated with the style code, and generating a 3D model of the object based on the output 3D shape.

Neural style transfer in three-dimensional shapes

One embodiment of the present invention sets forth a technique for training a machine learning model to perform style transfer. The technique includes applying one or more augmentations to a first input three-dimensional (3D) shape to generate a second input 3D shape. The technique also includes generating, via a first set of neural network layers, a style code based on a first latent representation of the first input 3D shape and a second latent representation of the second input 3D shape. The technique further includes generating, via a second set of neural network layers, a first output 3D shape based on the style code and the second latent representation, and performing one or more operations on the first and second sets of neural network layers based on a first loss associated with the first output 3D shape to generate a trained machine learning model.

Training machine learning models to perform neural style transfer in three-dimensional shapes

Methods, systems, and apparatus, including medium-encoded computer program products, for computer aided design and manufacture of physical structures include, in at least one aspect, a method including: obtaining a design space for a modeled object, load cases for physical simulation, and design criteria, wherein the modeled object includes specified geometry with which generatively designed geometry will connect, and wherein the load cases include at least one in-use load case for the physical structure and at least one subtractive-manufacturing load case associated with the specified geometry and with a subtractive manufacturing system; producing the generatively designed geometry in the design space for the modelled object in accordance with the load cases for physical simulation of the modelled object and the design criteria for the modeled object; and providing the modeled object with the generatively designed geometry for use in manufacturing the physical structure.

3D geometry generation for computer aided design considering subtractive manufacturing forces

In various embodiments, a workflow application generates and evaluates designs that reflect stylistic preferences. In operation, the workflow application determines a target style based on input received via a graphical user interface (GUI). Notably, the target style characterizes a first set of designs. The workflow application then generates stylized design(s) based on stylization algorithm(s) associated with the target style. Subsequently, the workflow application, displays a subset of the stylized design(s) via the GUI. A stylized design included in the subset of stylized design(s) is ultimately selected for production via the GUI. Advantageously, because the workflow application can substantially increase the number of designs that can be generated and evaluated based on the target style in a given amount of time, relative to more manual prior art techniques, the overall quality of the stylized design selected for production can be improved.

Techniques for generating designs that reflect stylistic preferences

Methods, systems, and apparatus, including medium-encoded computer program products, for computer aided design of physical structures using generative design processes, where three dimensional (3D) models of the physical structures can be produced to include lattices, hollows, holes, and combinations thereof, include: obtaining design criteria for an object; iteratively modifying 3D topology and shape(s) for the object using generative design process(es) that employ a macrostructure representation, e.g., using level-set method(s), in combination with physical simulation(s) that place void(s) in solid region(s) or solid(s) in void region(s) of the generative model of the object; and providing a 3D model of the generative design for the object for use in manufacturing a physical structure corresponding to the object using one or more computer-controlled manufacturing systems. The providing can include generating instructions for manufacturing machine(s), which can employ various manufacturing systems and techniques, including additive, subtractive and casting manufacturing methods.

Macrostructure topology generation with disparate physical simulation for Computer Aided Design and Manufacturing

Methods, systems, and apparatus, including medium-encoded computer program products, for designing three dimensional lattice structures include, in one aspect, a method including: obtaining a mechanical problem definition including a 3D model of an object; generating a numerical simulation model for the 3D model of the object using one or more loading cases and one or more isotropic solid materials identified as a baseline material model for a design space; predicting performance of different lattice settings in different orientations in the design space using a lattice structural behavior model in place of the baseline material model in the numerical simulation model; and presenting a set of lattice proposals for the design space based on the predicted performance of the different lattice settings in the different orientations; wherein the lattice structural behavior model has been precomputed for the different lattice settings, which are generable by the 3D modeling program.

Methods and systems for generating lattice recommendations in computer-aided design applications

In various embodiments, a stylization application generates designs that reflect stylistic preferences. In operation, the stylization application computes characterization information based on a first design and a trained machine-learning model that maps one or more designs to characterization information associated with one or more styles. The stylization application then computes a style score based on the characterization information and a target style that is included in the one or more styles. Subsequently, the stylization application generates a second design based on the style score, where the second design is more representative of the target style than the first design. Advantageously, because the stylization application can substantially increase the number of designs that can be generated based on the target style in a given amount of time, relative to more manual prior art techniques, the overall quality of the design ultimately selected for production can be improved.

In various embodiments, a stylization subsystem automatically modifies a three-dimensional (3D) object design. In operation, the stylization subsystem generates a simplified quad mesh based on an input triangle mesh that represents the 3D object design, a preferred orientation associated with at least a portion of the input triangle mesh, and mesh complexity constraint(s). The stylization subsystem then converts the simplified quad mesh to a simplified T-spline. Subsequently, the stylization subsystem creases one or more of edges included in the simplified T-spline to generate a stylized T-spline. Notably, the stylized T-spline represents a stylized design that is more convergent with the preferred orientation(s) than the 3D object design. Advantageously, relative to prior art approaches, the stylization subsystem can more efficiently modify the 3D object design to improve overall aesthetics and manufacturability.

Techniques for generating stylized quad-meshes from tri-meshes

Methods, systems, and apparatus, including medium-encoded computer program products, for computer aided design of physical structures using data format conversion (e.g., of output(s) from generative design processes) and user interface techniques that facilitate the production of 3D models of physical structures that are readily usable with 2.5-axis subtractive manufacturing, include: modifying smooth curves, which have been fit to contours representing discrete height layers of an object, to facilitate the 2.5-axis subtractive manufacturing; preparing an editable model of the object using a parametric feature history, which includes a sketch feature, to combine extruded versions of the smooth curves to form a 3D model of the object in a boundary representation format; reshaping a subset of the smooth curves responsive to user input with respect to the sketch feature; and replaying the parametric feature history to reconstruct the 3D model of the object, as changed by the user input.

Conversion of geometry to boundary representation with facilitated editing for computer aided design and 2.5-axis subtractive manufacturing

Growing community of inventors

Longfield, United Kingdom

Hooman Shayani