US 11,727,717 B2
Data-driven, photorealistic social face-trait encoding, prediction, and manipulation using deep neural networks
Alexander T. Todorov, Princeton, NJ (US); Stefan D. Uddenberg, Plainsboro, NJ (US); Joshua C. Peterson, Princeton, NJ (US); Thomas L. Griffiths, Princeton, NJ (US); and Jordan W. Suchow, New York, NY (US)
Assigned to THE TRUSTEES OF PRINCETON UNIVERSITY, Princeton, NJ (US); and THE TRUSTEES OF THE STEVENS INSTITUTE OF TECHNOLOGY, Hoboken, NJ (US)
Filed by THE TRUSTEES OF PRINCETON UNIVERSITY, Princeton, NJ (US); and THE TRUSTEES OF THE STEVENS INSTITUTE OF TECHNOLOGY, Hoboken, NJ (US)
Filed on Jan. 4, 2022, as Appl. No. 17/568,157.
Application 17/568,157 is a continuation of application No. 17/026,797, filed on Sep. 21, 2020, granted, now 11,250,245.
Claims priority of provisional application 62/903,267, filed on Sep. 20, 2019.
Prior Publication US 2022/0122378 A1, Apr. 21, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06V 40/16 (2022.01); G06N 3/04 (2023.01); G06V 10/764 (2022.01); G06V 10/774 (2022.01); G06V 10/82 (2022.01)
CPC G06V 40/165 (2022.01) [G06N 3/04 (2013.01); G06V 10/764 (2022.01); G06V 10/774 (2022.01); G06V 10/82 (2022.01); G06V 40/171 (2022.01); G06V 40/174 (2022.01)] 18 Claims
OG exemplary drawing
 
1. A system for photorealistic social face-trait encoding, prediction, and manipulation, comprising one or more processors configured with machine-readable instructions, that when executed cause the one or more processors to:
a. encode an image of a face as a multi-dimensional vector comprising one or more learned image features using an encoding process involving one or more stages;
b. modify the multi-dimensional vector to adjust at least one objective or subjective trait based on a learned function between the at least one objective or subjective trait and the one or more learned image features in the multi-dimensional vector; and
c. decode the modified multi-dimensional vector to generate an image of a realistic synthetic face.