US 12,192,543 B2
Video frame action detection using gated history
Gaurav Mittal, Redmond, WA (US); Ye Yu, Redmond, WA (US); Mei Chen, Bellevue, WA (US); and Junwen Chen, Rochester, NY (US)
Assigned to Microsoft Technology Licensing, LLC., Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Dec. 21, 2023, as Appl. No. 18/393,664.
Application 18/393,664 is a continuation of application No. 17/852,310, filed on Jun. 28, 2022, granted, now 11,895,343.
Claims priority of provisional application 63/348,993, filed on Jun. 3, 2022.
Prior Publication US 2024/0244279 A1, Jul. 18, 2024
Int. Cl. H04N 21/23 (2011.01); G06T 7/246 (2017.01); G06V 20/40 (2022.01); H04N 21/234 (2011.01)
CPC H04N 21/23418 (2013.01) [G06T 7/246 (2017.01); G06V 20/46 (2022.01); G06T 2207/10021 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
a processor; and
a computer-readable medium storing instructions that are operative upon execution by the processor to:
receive a video stream comprising a plurality of video frames;
group the plurality of video frames into a set of present video frames and a set of historical video frames, the set of present video frames comprising a current video frame;
based on at least the set of historical video frames and the set of present video frames, generate an action prediction for the current video frame; and
perform background suppression, wherein the action prediction comprises a confidence and wherein performing the background suppression comprises:
generating a loss function that weights low confidence video frames more heavily, with separate emphasis on action and background classes, for a classifier that generates the action prediction.