US 12,192,582 B2
System and method for removing copyrighted material from a streaming platform
Justin Sanford, San Francisco, CA (US); Matt Ward, San Francisco, CA (US); Salar Khan, San Francisco, CA (US); and Mark Franceschini, San Francisco, CA (US)
Assigned to Mux, Inc., San Francisco, CA (US)
Filed by Mux, Inc., San Francisco, CA (US)
Filed on Dec. 16, 2022, as Appl. No. 18/083,220.
Claims priority of provisional application 63/290,575, filed on Dec. 16, 2021.
Prior Publication US 2023/0199255 A1, Jun. 22, 2023
Int. Cl. H04N 21/4627 (2011.01); G06V 10/764 (2022.01); H04N 21/454 (2011.01)
CPC H04N 21/4627 (2013.01) [G06V 10/764 (2022.01); H04N 21/454 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising, for a first streaming event at a streaming platform:
during a first time period:
accessing a set of content logs for a population of content distribution networks including a set of streaming metrics for a video stream;
identifying an asset associated with the video stream;
accessing an asset profile for the asset, the asset profile including a set of profile parameters; and
computing a risk score for the asset according to a score range and based on a subset of profile parameters; and
during a second time period:
monitoring the set of streaming metrics for the video stream during a set of time intervals of a first duration during a first time window;
in response to a first streaming metric, in the set of streaming metrics, for the first video stream exceeding a threshold metric during the first time window, accessing a conflict log including a set of flagged assets;
in response to identifying absence of the asset in the conflict log:
retrieving the risk score for the asset;
extracting a set of image frames from the video stream according to an image extraction frequency;
executing an image classification model based on the set of image frames to characterize the image frames according to a set of tags;
retrieving a content manifest associated with a content type of the video stream, the content manifest defining a set of target concepts related to the content type;
deriving a difference between the set of tags associated with the set of image frames to the set of target concepts in the content manifest to compute a match score for the video stream;
in response to the match score exceeding a threshold score:
flagging the video stream for manual authentication; and
delivering the video stream to an interface associated with an operator; and
in response to receiving an abuse confirmation for the video stream from the operator:
generating a notification n descriptive of the abuse confirmation;
transmitting the notification to the asset; and
performing an action on the asset profile based on the risk score for the asset; and
in response to identifying the asset in the conflict log:
identifying the asset profile associated with the asset; and
automatically deactivating the asset profile to prevent the first streaming event of the video stream via the streaming platform.