Granted Patent B1

US 11,217,036 · App. 16/595,285 · Granted Jan 4, 2022

Inventors: Elif Albuz (Los Gatos, CA); Chad Vernon (South San Francisco, CA); Shu Liang (Redwood City, CA); Peihong Guo (San Mateo, CA)

Assignee: Facebook Technologies, LLC

G06T19/20G06K9/00268G06T2219/2012G06T2219/2021

Abstract

Claims (58)

1. A method for generating a personalized avatar, the method comprising:

obtaining user biographic data;

creating a user model based on the user biographic data;

obtaining an average person model with dimensions that are based on averages of measurements from multiple people;

computing a delta between the user model and the average person model, by:

computing one or more differences, for a given facial feature, between the user model for that given facial feature and the average person model for that given facial feature;

obtaining a generic avatar model, different from the average person model, that represents a style in which the personalized avatar is to be created; and

modifying the generic avatar model, different from the average person model, to create the personalized avatar in the style by:

distinguishing multiple facial features of the generic avatar model having corresponding amounts, identified in the delta, specifying how the user model differs from the average person model for that distinguished facial feature; and

adjusting each of the multiple distinguished facial features of the generic avatar model according to the corresponding amount, identified in the delta.

2. The method of claim 1 , wherein the user biographic data comprises one or more images, each depicting at least part of a user.

3. The method of claim 1 , wherein the user model comprises one or both of:

a set of measurements corresponding to dimensions or curvatures of portions of a user; or

a 3D model of at least part of the user.

4. The method of claim 3 , wherein creating the user model comprises:

identifying measurements of the user's face including one or more curvatures of at least one facial feature, observable in one or more images from the user biographic data; and

creating a 3D model with dimensions of facial features matching the identified measurements.

5. The method of claim 4 , wherein identifying measurements of the user's face comprises:

estimating one or more angles between the user and a camera that captured the one or more images; and

using the estimates to determine measurements of facial features shown in the one or more images.

6. The method of claim 3 , wherein creating a user model comprises applying a machine learning model that receives image data and is trained to identify measurements of user features based on the image data.

7. The method of claim 1 further comprising identifying shading, textures, or styles for identified portions of the user by identifying areas in the images corresponding to facial features and mapping the shading, textures, or styles from those areas to corresponding user parts.

8. The method of claim 1 further comprising:

identifying additional user characteristics comprising one or more of: hair color, hairstyle, eye color, eyebrow geometry, glasses style, piercing configuration, facial hair configuration, or any combination thereof;

selecting, from a library, one or more additional features that correspond to the identified additional user characteristics; and

applying, to the created personalized avatar, the selected one or more additional features.

9. The method of claim 8 , wherein the library is a genre-specific library associated with the generic avatar model.

10. The method of claim 1 , wherein the average person model comprises a 3D model.

11. The method of claim 1 , wherein the measurements from multiple people are from multiple people that share one or more specified characteristics with the user including one or more of: gender, ethnicity, an age range, or any combination thereof.

12. The method of claim 1 , wherein the delta comprises at least one specification of a difference in curvatures between a part of a face of the user and a part of a facial representation from the average person model.

13. A computer-readable storage medium storing instructions that, when executed by a computing system, cause the computing system to perform operations for generating a personalized avatar, the operations comprising:

obtaining an average person model that is based on measurements from multiple people;

computing a delta between A) a user model that is based on one or more user images, and B) the average person model, by:

computing one or more differences, for a given feature, between the user model for that given feature and the average person model for that given feature;

obtaining a generic avatar model, different from the average person model, that represents a style in which the personalized avatar is to be created; and

modifying the generic avatar model, different from the average person model, to create the personalized avatar in the style by:

distinguishing multiple facial features of the generic avatar model having corresponding amounts, identified in the delta, specifying how the user model differs from the average person model for that distinguished facial feature; and

adjusting each of the multiple distinguished facial features of the generic avatar model according to the corresponding amount, identified in the delta.

14. The computer-readable storage medium of claim 13 , wherein the user model is created by:

identifying measurements of the user's face including one or more curvatures of at least one facial feature, observable in the one or more user images; and

creating a 3D model with dimensions of facial features matching the identified measurements.

15. The computer-readable storage medium of claim 13 , wherein the measurements from multiple people are from multiple people that share a set of one or more specified characteristics with a user depicted in the one or more user images, the set including one or more of: gender, ethnicity, an age range, or any combination thereof.

16. The computer-readable storage medium of claim 13 , wherein the delta comprises at least one specification of a difference in curvatures between part of a face of the user and a part of a facial representation from the average person model.

17. A computing system for generating a personalized avatar, the computing system comprising:

one or more processors; and

a memory storing instructions that, when executed by the computing system, cause the computing system to perform operations comprising:

creating a user model based on one or more user images;

computing a delta between the user model and an average person model, by:

computing one or more differences, for a given feature, between the user model for that given feature and the average person model for that given feature; and

obtaining a generic avatar model, different from the average person model, that represents a style in which the personalized avatar is to be created; and

modifying the generic avatar model, different from the average person model, to create the personalized avatar in the style by:

distinguishing multiple facial features of the generic avatar model having corresponding amounts, identified in the delta, specifying how the user model differs from the average person model for that distinguished facial feature; and

adjusting each of the multiple distinguished facial features of the generic avatar model according to the corresponding amount, identified in the delta.

18. The computing system of claim 17 , wherein the an average person model is based on measurements from multiple people that are from multiple people that share a set of one or more specified characteristics with a user depicted in the one or more user images, the set including one or more of: gender, ethnicity, an age range, or any combination thereof.

19. The computing system of claim 17 , wherein the delta comprises at least one specification of a difference in curvatures between part of a face of the user and a part of a facial representation from the average person model.

20. The computing system of claim 17 , wherein the adjusting at least one of the multiple distinguished facial features of the generic avatar model comprises:

identifying, in the delta, a percentage difference between the user model and the average person model for the at least one distinguished facial feature; and

modifying the corresponding at least one distinguished facial feature of the generic avatar model by the percentage difference.

Assignments (2)

Cited By (10)

US 12,217,347 US 12,277,661 US 12,326,568 US 12,340,481 US 12,379,834 US 12,417,596 US 12,444,114 US 12,536,739 US 12,561,930 US 12,713,125