US 12,273,604 B2
Method and system to highlight video segments in a video stream
Nachiketa Mishra, Fremont, CA (US); Steven Roeder, Seattle, WA (US); and Stephen Slotnick, Seattle, WA (US)
Assigned to iStreamPlanet Co., LLC, Las Vegas, NV (US)
Filed by iStreamPlanet Co., LLC, Las Vegas, NV (US)
Filed on Mar. 16, 2022, as Appl. No. 17/696,747.
Prior Publication US 2023/0300430 A1, Sep. 21, 2023
Int. Cl. H04N 21/8549 (2011.01); G06V 10/82 (2022.01); G11B 27/34 (2006.01); H04N 21/234 (2011.01); H04N 21/431 (2011.01)
CPC H04N 21/8549 (2013.01) [G06V 10/82 (2022.01); G11B 27/34 (2013.01); H04N 21/23418 (2013.01); H04N 21/4312 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A method executed by an electronic device in a video streaming platform, the method comprising:
receiving, a media stream from a media source, wherein the media stream includes a video stream and an audio stream;
extracting, at least one feature from the video stream and the audio stream;
determining a context corresponding to the at least one feature, wherein the context includes a source type or a source event;
identifying, via a machine learning model, a highlight segment within the video stream based on the context, the identifying including determining a starting frame and an ending frame of the highlight segment, wherein the starting frame and the ending frame of the highlight segment are identified by inputting the video stream and audio data corresponding to the at least one feature into the machine-learning model;
stitching one or more individual frames corresponding to the at least one feature between the starting frame and the ending frame of the highlight segment, wherein at least one frame of the one or more individual frames is a non-consecutive frame; and
indicating an availability of the stitched one or more individual frames to a playback device in the video stream, wherein indicating the availability includes providing a confidence score based on the context, wherein the indication is aligned with timing data of the video stream.