Techniques for generating designs that reflect stylistic preferences

Patent metadata

Field	Value
Publication number	US-11977960-B2
Application number	US-201916534985-A
Country	US
Kind code	B2
Filing date	Aug 7, 2019
Priority date	Aug 9, 2018
Publication date	May 7, 2024
Grant date	May 7, 2024

Abstract

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.

First claim

What is claimed is:

1. A computer-implemented method for generating and evaluating designs that accounts for stylistic preferences, the method comprising:
determining a first style and a second style based on a first input and a second input received via a graphical user interface (GUI), wherein the first style characterizes a first plurality of designs and the second style characterizes a second plurality of designs;
generating one or more stylized designs based on one or more stylization algorithms associated with the first style and the second style; and
displaying, in a graphical plot via the GUI, a first set of data points representing a first subset of the one or more stylized designs associated with the first style and a second set of data points representing a second subset of the one or more stylized designs associated with the second style, wherein the first set of data points is displayed with a first visual appearance, wherein the second set of data points is displayed with a second visual appearance that is different from the first visual appearance, and wherein the first set of data points and the second set of data points are displayed along a first axis of the graphical plot representing a first performance metric of the one or more stylized designs, wherein the first performance metric represents a functional characteristic of the one or more stylized designs.

2. The computer-implemented method of claim 1 , wherein generating the one or more stylized designs comprises executing an optimization algorithm that modifies an initial design based on the first style and a first trained machine-learning model that encapsulates the one or more stylization algorithms.

3. The computer-implemented method of claim 1 , wherein generating the one or more stylized designs comprises executing a generative design algorithm based on the first style and a first trained machine-learning model that encapsulates the one or more stylization algorithms.

4. The computer-implemented method of claim 3 , further comprising performing one or more machine-learning operations based on the first plurality of designs to generate the first trained machine-learning model.

5. The computer-implemented method of claim 1 , wherein generating the one or more stylized designs comprises:
determining a design construction set that encapsulates the one or more stylization algorithms based on the first style, wherein the design construction set includes at least one of a design primitive, a designs element, and a design operation; and
constructing the one or more stylized designs based on the design construction set.

6. The computer-implemented method of claim 1 , wherein a first trained machine-learning model encapsulates the one or more stylization algorithms, wherein one or more re-training operations are performed on the first trained machine-learning model to generate a second trained machine-learning model, wherein a first accuracy associated with the first trained machine-learning model is lower than a second accuracy associated with the second trained machine-learning model.

7. The computer-implemented method of claim 1 , further comprising:
computing one or more style scores based on the first style, the one or more stylized designs, and a first trained machine-learning model that maps one or more designs to characterization information associated with one or more styles; and
performing at least one of a sorting operation, a filtering operation, and a clustering operation on the one or more stylized designs using the one or more style scores to generate the first subset of the one or more stylized designs.

8. The computer-implemented method of claim 1 , wherein a first stylized design included in the first subset of the one or more stylized designs comprises a computer-aided design (CAD) geometry model, a point cloud, or a three- dimensional (3D) image.

9. The computer-implemented method of claim 1 , wherein the first style is characterized by at least one of an aesthetic trait and a manufacturing-related property.

10. One or more non-transitory computer readable media including instructions that, when executed by one or more processors, cause the one or more processors to generate and evaluate designs that account for stylistic preferences by performing the steps of:
determining a first style and a second style based on a first input and a second input received via a graphical user interface (GUI), wherein the first style characterizes a first plurality of designs and the second style characterizes a second plurality of designs;
generating one or more stylized designs based on one or more stylization algorithms associated with the first style and the second style; and
displaying, in a graphical plot via the GUI, a first set of data points representing a first subset of the one or more stylized designs associated with the first style and a second set of data points representing a second subset of the one or more stylized designs associated with the second style, wherein the first set of data points is displayed with a first visual appearance, wherein the second set of data points is displayed with a second visual appearance that is different from the first visual appearance, and wherein the first set of data points and the second set of data points are displayed along a first axis of the graphical plot representing a first performance metric of the one or more stylized designs, wherein the first performance metric represents a functional characteristic of the one or more stylized designs.

11. The one or more non-transitory computer readable media of claim 10 , wherein generating the one or more stylized designs comprises executing an optimization algorithm that modifies an initial design based on the first style and a first trained machine-learning model that encapsulates the one or more stylization algorithms.

12. The one or more non-transitory computer readable media of claim 10 , wherein generating the one or more stylized designs comprises executing a generative design algorithm based on the first style and a first trained machine-learning model that encapsulates the one or more stylization algorithms.

13. The one or more non-transitory computer readable media of claim 12 , further comprising performing one or more machine-learning operations based on the first plurality of designs to generate the first trained machine-learning model.

14. The one or more non-transitory computer readable media of claim 10 , wherein generating the one or more stylized designs comprises:
determining that the first style is a positive target based on the first input;
determining a design construction set that encapsulates the one or more stylization algorithms based on the first style, wherein the design construction set includes at least one of a design primitive, a designs element, and a design operation; and
constructing the one or more stylized designs based on the design construction set.

15. The one or more non-transitory computer readable media of claim 10 , wherein a first trained machine-learning model encapsulates the one or more stylization algorithms, wherein one or more re-training operations are performed on the first trained machine-learning model to generate a second trained machine-learning model, wherein a first accuracy associated with the first trained machine-learning model is lower than a second accuracy associated with the second trained machine-learning model.

16. The one or more non-transitory computer readable media of claim 10 , further comprising:
computing one or more style scores based on the first sty

Assignees

• Autodesk Inc

Inventors

• Shayani Hooman
• Davis Mark Thomas

Classifications

• G06T11/10

  Texturing; Colouring; Generation of textures or colours (retouching, inpainting or scratch removal G06T5/77) · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06F30/00

  Computer-aided design [CAD] · CPC title

• G06F30/17

  Mechanical parametric or variational design · CPC title