US 12,078,730 B2
Single click box prediction for temporal Lidar labeling automation
Meng-Ta Chou, San Francisco, CA (US); Jennifer Villa, Palo Alto, CA (US); Matt Deiters, San Francisco, CA (US); Mesut Arik, San Francisco, CA (US); Radu Dondera, Oakland, CA (US); and Yunjing Xu, San Francisco, CA (US)
Assigned to GM Cruise Holdings LLC, San Francisco, CA (US)
Filed by GM Cruise Holdings LLC, San Francisco, CA (US)
Filed on Mar. 31, 2021, as Appl. No. 17/219,731.
Prior Publication US 2022/0317305 A1, Oct. 6, 2022
Int. Cl. G01S 17/931 (2020.01); B60W 60/00 (2020.01); G01S 7/48 (2006.01); G01S 17/04 (2020.01); G01S 17/42 (2006.01); G05D 1/00 (2006.01)
CPC G01S 17/931 (2020.01) [B60W 60/0027 (2020.02); G01S 7/4808 (2013.01); G01S 17/04 (2020.01); G01S 17/42 (2013.01); G05D 1/0251 (2013.01); G05D 1/0276 (2013.01); B60W 2420/408 (2024.01); B60W 2540/215 (2020.02); B60W 2554/404 (2020.02)] 17 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving, at a labeling automation system, point cloud data based on a detected object from an autonomous vehicle;
determining, by the labeling automation system, an object based on the point cloud data of the detected object and corresponding models related to the detected object;
updating, by the labeling automation system, the models based on the point cloud data of the detected object and selected object labels, the selected object labels being based on 3D bounding boxes of the detected object;
providing, by the labeling automation system, the updated models to the autonomous vehicle for deployment; and
providing, by the labeling automation system, the 3D bounding boxes and a selection from a labeling user interface to a heuristic tracking system of labeling automation system.