Back to results
US 12,738,079 B1Granted
Neural network system for generating video of objects outside training set
AssigneeNVIDIA Corporation
InventorsTing-Chun Wang, Bryan Christopher Catanzaro, Guilin Liu, Ming-Yu Liu, Andrew James Tao
Life of this patent
Estimating the term…
Abstract
Apparatuses, systems, and techniques to generate video of objects. In at least one embodiment, one or more neural networks generate video of objects whose images are different from those in training sets of these neural networks.
Claims50
One or more processors, comprising:
- circuitry to:
- cause one or more first neural networks to calculate one or more parameters of one or more second neural networks, wherein the one or more parameters are calculated based, at least in part, on one or more sample images of one or more objects;
- cause the one or more parameters of the one or more second neural networks to be updated based on at least one or more characteristics of the one or more objects;
- and
- generate one or more video frames, using the one or more second neural networks with the one or more updated parameters based, at least in part, on one or more semantic images corresponding to the one or more sample images, wherein the one or more semantic images comprise one or more encodings of geometric information of the one or more objects.
2 more elements
A system comprising:
- one or more memories to store one or more first neural networks and one or more second neural networks;
- and
- one or more processors to cause one or more first neural networks to calculate one or more parameters of one or more second neural networks, wherein the one or more parameters are calculated based, at least in part, on one or more sample images of one or more objects;
- cause the one or more parameters of the one or more second neural networks to be updated based on at least one or more characteristics of the one or more objects;
- and
- generate one or more video frames, using the one or more second neural networks with the one or more updated parameters, based, at least in part, on one or more semantic images corresponding to the one or more sample images, wherein the one or more semantic images comprise one or more encodings of geometric information of the one or more objects.
3 more elements
A system comprising:
- one or more memories to store one or more first images of one or more first objects;
- and
- one or more processors to cause one or more first neural networks to calculate one or more parameters of one or more second neural networks, wherein the one or more parameters are calculated based, at least in part, on one or more sample images of one or more objects;
- cause the one or more parameters to be updated based on at least one or more characteristics of the one or more objects;
- and
- generate one or more video frames, using the one or more second neural networks with the one or more updated parameters based, at least in part, on one or more semantic images corresponding to the one or more sample images, wherein the one or more semantic images comprise one or more encodings of geometric information of the one or more objects.
3 more elements
One or more processors comprising:
- circuitry to train one or more first neural networks to calculate one or more parameters of one or more second neural networks, wherein the one or more parameters are calculated based, at least in part, on one or more sample images of one or more objects;
- cause the one or more parameters of the one or more second neural networks to be updated based on at least one or more characteristics of the one or more objects;
- and
- generate one or more video frames, using the one or more second neural networks with the one or more updated parameters based, at least in part, on one or more semantic images corresponding to the one or more sample images, wherein the one or more semantic images comprise one or more encodings of geometric information of the one or more objects.
1 more element
A video gaming platform comprising:
- one or more memories to store a digital representation of one or more first images corresponding to a video game, wherein the one or more first images are images of one or more first objects;
- and
- one or more processors to cause one or more first neural networks to calculate one or more parameters of one or more second neural networks, wherein the one or more parameters are calculated based, at least in part, on one or more sample images of one or more objects;
- cause the one or more parameters of the one or more second neural networks to be updated based on at least one or more characteristics of the one or more objects;
- and
- generate one or more video frames, using the one or more second neural networks with the one or more updated parameters based, at least in part, on one or more semantic images corresponding to the one or more sample images, wherein the one or more semantic images comprise one or more encodings of geometric information of the one or more objects.
3 more elements
A method, comprising:
- causing one or more first neural networks to
- calculate one or more parameters of one or more second neural networks, wherein the one or more parameters are calculated based, at least in part, on one or more sample images of one or more objects;
- cause the one or more parameters of the one or more second neural networks to be updated based on at least one or more characteristics of the one or more objects;
- and
- generate one or more video frames, using the one or more second neural networks with the one or more updated parameters based, at least in part, on one or more semantic images corresponding to the one or more sample images, wherein the one or more semantic images comprise one or more encodings of geometric information of the one or more objects.
2 more elements
Dependencies are inferred from the claim text.