US 12,461,961 B2
Generating and utilizing digital media clips based on contextual metadata from digital environments
Arunsunai Anbukarasi Anbalagapandian, San Francisco, CA (US); Devin Mancuso, Alameda, CA (US); Rituparna Vincent, Alamo, CA (US); and Viksit Gaur, San Francisco, CA (US)
Assigned to Dropbox, Inc., San Francisco, CA (US)
Filed by Dropbox, Inc., San Francisco, CA (US)
Filed on Sep. 18, 2023, as Appl. No. 18/469,319.
Application 18/469,319 is a continuation of application No. 17/657,572, filed on Mar. 31, 2022, granted, now 11,762,898.
Prior Publication US 2024/0004916 A1, Jan. 4, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 16/435 (2019.01); G06F 16/44 (2019.01); G06F 16/45 (2019.01); G06F 16/48 (2019.01)
CPC G06F 16/435 (2019.01) [G06F 16/44 (2019.01); G06F 16/45 (2019.01); G06F 16/48 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
maintaining, for a user account of a content management system, a media clip library comprising digital media clips each comprising a content item, content metadata, and contextual metadata;
identifying a digital media clip from the digital media clips based on receiving an indication of a user selection of the digital media clip;
determining a relatedness level based on receiving an indication of a user selection corresponding to a similarity metric;
generating a search request based on the identified digital media clip and the relatedness level;
tuning a media clip classification machine-learning model based on the relatedness level and the content metadata and the contextual metadata associated with the digital media clips of the user account;
identifying search results comprising one or more content items from one or more sources utilizing the tuned media clip classification machine-learning model and the identified digital media clip of the search request; and
providing the search results comprising the one or more content items in response to the search request.