US 12,406,471 B2
Using temporal filters for automated real-time classification
Sakthivel Sivaraman, Santa Clara, CA (US); Shagan Sah, Santa Clara, CA (US); and Niranjan Avadhanam, Saratoga, CA (US)
Assigned to NVIDIA Corporation, Santa Clara, CA (US)
Filed by NVIDIA Corporation, Santa Clara, CA (US)
Filed on Jul. 12, 2024, as Appl. No. 18/771,646.
Application 18/771,646 is a continuation of application No. 18/333,281, filed on Jun. 12, 2023, granted, now 12,073,604.
Application 18/333,281 is a continuation of application No. 17/647,390, filed on Jan. 7, 2022, granted, now 11,721,089, issued on Aug. 8, 2023.
Application 17/647,390 is a continuation of application No. 16/907,125, filed on Jun. 19, 2020, granted, now 11,222,232, issued on Jan. 11, 2022.
Prior Publication US 2024/0371136 A1, Nov. 7, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06V 10/00 (2022.01); G06F 18/20 (2023.01); G06F 18/2321 (2023.01); G06N 3/045 (2023.01); G06V 10/764 (2022.01); G06V 20/40 (2022.01); G06V 40/10 (2022.01); G06V 40/20 (2022.01)
CPC G06V 10/764 (2022.01) [G06F 18/2321 (2023.01); G06F 18/285 (2023.01); G06N 3/045 (2023.01); G06V 20/47 (2022.01); G06V 40/113 (2022.01); G06V 40/28 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
determining, using one or more neural networks and based at least on sensor data obtained using one or more sensors of a machine, a temporal series of classification outputs;
generating one or more filtered classification outputs based at least on applying a filter to one or more outputs of the temporal series of classification outputs;
determining one or more confidence scores for one or more frames of the sensor data using the one or more filtered classification outputs; and
performing one or more operations associated with the machine based at least on the one or more confidence scores.