	US 12,472,628 B2
	Grasp pose prediction
Jonathan Tremblay, Redmond, WA (US); Stanley Thomas Birchfield, Sammamish, WA (US); Valts Blukis, Seattle, WA (US); Bowen Wen, Bellevue, WA (US); Dieter Fox, Seattle, WA (US); and Taeyeop Lee, Daejeon (KR)
Assigned to NVIDIA Corporation, Santa Clara, CA (US)
Filed by NVIDIA Corporation, Santa Clara, CA (US)
Filed on Jul. 6, 2023, as Appl. No. 18/219,031.
Claims priority of provisional application 63/411,486, filed on Sep. 29, 2022.
Prior Publication US 2024/0123620 A1, Apr. 18, 2024
Int. Cl. B25J 9/16 (2006.01)

CPC B25J 9/1669 (2013.01) [B25J 9/161 (2013.01); B25J 9/1697 (2013.01)]

20 Claims

1. A system, comprising:

at least one processor; and

at least one memory comprising instructions that, in response to execution by the at least one processor, cause the system to at least:

obtain an input image depicting an object;

generate a first set of images by at least using a first neural network to process a latent code corresponding to the object;

update the latent code based, at least in part, on the first set of images and the input image;

use a second neural network to generate a set of grasp proposals based, at least in part, on the updated latent code; and

select one or more grasp proposals from the set of grasp proposals based, at least in part, on the first neural network and the updated latent code.