US 12,462,377 B1
Object detection using one or more neural networks
Holger Roth, Rockville, MD (US); Ling Zhang, Rockville, MD (US); Dong Yang, North Bethesda, MD (US); Fausto Milletari, Munich (DE); Ziyue Xu, Reston, VA (US); Xiaosong Wang, Rockville, MD (US); and Daguang Xu, Potomac, MD (US)
Assigned to NVIDIA Corporation, Santa Clara, CA (US)
Filed by Nvidia Corporation, Santa Clara, CA (US)
Filed on Aug. 14, 2019, as Appl. No. 16/540,694.
Int. Cl. G06T 7/00 (2017.01); G06N 3/045 (2023.01); G06N 3/08 (2023.01); G06T 7/11 (2017.01)
CPC G06T 7/0012 (2013.01) [G06N 3/045 (2023.01); G06N 3/08 (2013.01); G06T 7/11 (2017.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01)] 32 Claims
OG exemplary drawing
 
1. One or more processors, comprising:
circuitry to:
cause one or more neural networks to generate a first segmentation of an object depicted in an image based, at least in part, on a plurality of pixels selected randomly from within a user-indicated boundary area that surrounds the object; and
train the one or more neural networks to generate a second segmentation of the object based, at least in part, on the first segmentation of the object, wherein the one or more neural networks are trained to generate the second segmentation based at least on use of the first segmentation as a supervisory signal.