RIG-SPACE NEURAL RENDERING OF DIGITAL ASSETS

Number: 20210233300
Published: 2021-07-29
Filed: 2020-03-05
Assignee: DISNEY ENTERPRISES, INC.
Inventors: BORER; Dominik, BUHMANN; Jakob, GUAY; Martin
CPC: G06N3/09; G06N3/0464; G06N20/00; G06N3/045; G06N3/0475; G06T13/40
Verdict: Low Notable software
Source: Google Patents · FreePatentsOnline

The keeper's note

Neural rendering technique for rigged digital assets.

Abstract

Various embodiments disclosed herein provide techniques for generating image data of a three-dimensional (3D) animatable asset. A rendering module executing on a computer system accesses a machine learning model that has been trained via first image data of the 3D animatable asset generated from first rig vector data. The rendering module receives second rig vector data. The rendering module generates, via the machine learning model, a second image data of the 3D animatable asset based on the second rig vector data.

Background

BACKGROUND Field of the Various Embodiments

The various embodiments relate generally to computer-based animation and, more specifically, to rig-space neural rendering of digital assets. Description of the Related Art

Oftentimes, animated and live-action movie productions for feature films, short subjects, and/or the like employ computer-based animation and/or computer-generated imagery (CGI). Typically, such movie productions are intended to be displayed on large-format displays, such as cinema screens. Accordingly, movie production professionals create high resolution three-dimensional (3D) characters and other synthetic 3D objects to generate high quality images capable of being projected onto such large displays. Such 3D characters and 3D objects are referred to herein as “3D assets.” These high-resolution 3D assets are animated by means of complex proprietary rigs, where a rig provides structure to the 3D asset that enable the 3D asset to move. Computer animators manipulate the rig to cause the 3D asset to move, using a process known as “deformation.” Surface geometry defines the look of the 3D asset. As the rig moves, the surface geometry moves correspondingly, resulting in a lifelike 3D asset that can move and interact with other assets in the movie production. Due to the complexity of high-resolution 3D assets developed for movie productions, these 3D assets are typically not animated in real-time. For example, each second of animation for a 3D asset may

Claims

1. A computer-implemented method for generating image data of a three-dimensional (3D) animatable asset, the method comprising: accessing a machine learning model that has been trained via first image data of the 3D animatable asset generated from first rig vector data; receiving second rig vector data; and generating, via the machine learning model, a second image data of the 3D animatable asset based on the second rig vector data. || 12. One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of: accessing a machine learning model that has been trained via first image data of the 3D animatable asset generated from first rig vector data; receiving second rig vector data; and generating, via the machine learning model, a second image data of the 3D animatable asset based on the second rig vector data. || 20. A system, comprising: a memory that includes instructions; and a processor that is coupled to the memory and, when executing the instructions: accesses a machine learning model that has been trained via first image data of the 3D animatable asset generated from first rig vector data; receives second rig vector data; and generates, via the machine learning model, a second image data of the 3D animatable asset based on the second rig vector data.