US 11,676,364 B2
Real-time detection of lanes and boundaries by autonomous vehicles
Yifang Xu, San Jose, CA (US); Xin Liu, Pleasanton, CA (US); Chia-Chih Chen, San Jose, CA (US); Carolina Parada, Boulder, CO (US); Davide Onofrio, San Francisco, CA (US); Minwoo Park, Cupertino, CA (US); Mehdi Sajjadi Mohammadabadi, Santa Clara, CA (US); Vijay Chintalapudi, Sunnyvale, CA (US); Ozan Tonkal, Munich (DE); John Zedlewski, San Francisco, CA (US); Pekka Janis, Uusimaa (FI); Jan Nikolaus Fritsch, Santa Clara, CA (US); Gordon Grigor, San Francisco, CA (US); Zuoguan Wang, Los Gatos, CA (US); I-Kuei Chen, Milpitas, CA (US); and Miguel Sainz, Palo Alto, CA (US)
Assigned to NVIDIA Corporation, Santa Clara, CA (US)
Filed by NVIDIA Corporation, Santa Clara, CA (US)
Filed on Apr. 5, 2021, as Appl. No. 17/222,680.
Application 17/222,680 is a continuation of application No. 16/286,329, filed on Feb. 26, 2019, granted, now 10,997,433.
Claims priority of provisional application 62/636,142, filed on Feb. 27, 2018.
Prior Publication US 2021/0224556 A1, Jul. 22, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G06V 10/44 (2022.01); G06T 7/10 (2017.01); G05D 1/00 (2006.01); G06N 3/084 (2023.01); G05D 1/02 (2020.01); G06V 20/56 (2022.01); G06V 10/46 (2022.01); G06V 20/40 (2022.01); G06F 18/2413 (2023.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01)
CPC G06V 10/44 (2022.01) [G05D 1/0088 (2013.01); G05D 1/0221 (2013.01); G06F 18/24143 (2023.01); G06N 3/084 (2013.01); G06T 7/10 (2017.01); G06V 10/457 (2022.01); G06V 10/46 (2022.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06V 20/41 (2022.01); G06V 20/588 (2022.01); G06V 10/471 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A processor comprising:
processing circuitry to:
compute, using a neural network and based at least on sensor data generated using one or more sensors of an ego-machine, a multi-class mask indicative of first points corresponding to a first road marking type and second points corresponding to a second road marking type;
compute, based at least on the multi-class mask, at least a first edge corresponding to the first road marking type using two or more of the first points and a second edge corresponding to the second road marking type using two or more of the second points;
assign a first label to the first edge and a second label to the second edge based at least on relative locations of the first edge and the second edge with respect to the ego-machine; and
perform one or more operations by the ego-machine based at least on the first edge, the second edge, the first label, and the second label.