US 11,961,188 B2
Neural networks to generate appearance-responsive material map sets in digital graphical environments
Robin Faury, Clermont-Ferrand (FR); Tamy Boubekeur, Paris (FR); Jeremy Levallois, Lyons (FR); Alban Gauthier, Paris (FR); and Theo Thonat, Issy les Moulineaux (FR)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by Adobe Inc., San Jose, CA (US)
Filed on Jun. 8, 2022, as Appl. No. 17/835,754.
Prior Publication US 2023/0401793 A1, Dec. 14, 2023
Int. Cl. G06T 17/30 (2006.01); G06T 7/00 (2017.01); G06T 15/04 (2011.01)
CPC G06T 17/30 (2013.01) [G06T 7/97 (2017.01); G06T 15/04 (2013.01); G06T 2200/04 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system for generating an appearance-responsive material map, the system comprising:
an appearance-responsive material map generation subsystem configured for:
accessing source material data, the source material data including material data that describes one or more surface appearance values associated with a three-dimensional (“3D”) graphical object;
identifying, via a trained material map generation neural network, a combination of features from the source material data, wherein the trained material map generation neural network is trained to select the combination of features based on a detail level;
generating an appearance-responsive material map that includes a combination of material data from the source material data, the combination of material data associated with the identified combination of features; and
providing the appearance-responsive material map to a rendering subsystem; and
the rendering subsystem, which is configured for:
selecting the appearance-responsive material map based on a viewpoint distance, and
rendering the 3D graphical object having a visual appearance that is based on the appearance-responsive material map.