US 12,033,078 B2
Sequence-of-sequences model for 3D object recognition
Yuncheng Li, Los Angeles, CA (US); Zhou Ren, Bellevue, WA (US); Ning Xu, Irvine, CA (US); Enxu Yan, Los Altos, CA (US); and Tan Yu, Singapore (SG)
Assigned to SNAP INC., Santa Monica, CA (US)
Filed by Snap Inc., Santa Monica, CA (US)
Filed on Aug. 4, 2023, as Appl. No. 18/230,499.
Application 18/230,499 is a continuation of application No. 17/878,591, filed on Aug. 1, 2022, granted, now 11,755,910.
Application 17/878,591 is a continuation of application No. 16/870,138, filed on May 8, 2020, granted, now 11,410,439.
Claims priority of provisional application 62/845,471, filed on May 9, 2019.
Prior Publication US 2023/0376757 A1, Nov. 23, 2023
Int. Cl. G06N 3/08 (2023.01); G06F 18/214 (2023.01); G06F 18/2431 (2023.01); G06T 7/33 (2017.01); G06T 7/55 (2017.01); G06V 10/44 (2022.01); G06V 10/82 (2022.01); G06V 20/20 (2022.01); G06V 20/40 (2022.01); G06V 20/64 (2022.01)
CPC G06N 3/08 (2013.01) [G06F 18/214 (2023.01); G06F 18/2431 (2023.01); G06T 7/344 (2017.01); G06T 7/55 (2017.01); G06V 10/454 (2022.01); G06V 10/82 (2022.01); G06V 20/20 (2022.01); G06V 20/41 (2022.01); G06V 20/64 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
accessing, from a plurality of virtual cameras, multiple sequences of views of a three-dimensional object, each sequence of the multiple sequences representing a unique order of views, wherein each sequence of views of the multiple sequences of views has a unique starting view;
generating, using one or more processors, aligned sequences from the multiple sequences based on an alignment function associated with the unique starting view of each sequence of views of the multiple sequences of views; and
identifying the three-dimensional object using the aligned sequences.