US 12,216,704 B2
Context-based organization of digital media
Rotish Kumar, Telangana (IN); Arshla Jindal, Nodia (IN); and Abhishek Shah, Delhi (IN)
Assigned to Adobe Inc., San Jose, CA (US)
Filed by ADOBE INC., San Jose, CA (US)
Filed on Oct. 3, 2019, as Appl. No. 16/591,913.
Prior Publication US 2021/0103611 A1, Apr. 8, 2021
Int. Cl. G06F 16/583 (2019.01); G06F 16/538 (2019.01); G06F 16/54 (2019.01); G06F 16/58 (2019.01); G06F 16/587 (2019.01)
CPC G06F 16/583 (2019.01) [G06F 16/538 (2019.01); G06F 16/54 (2019.01); G06F 16/5866 (2019.01); G06F 16/587 (2019.01)] 19 Claims
OG exemplary drawing
 
1. A computerized method comprising:
storing, in a data store, a representation of a plurality of event types, a set of learned content tags learned from a training dataset to be representative of each corresponding event type of the event types, and, for each learned content tag in the set for each corresponding event type, a learned correlation score that quantifies a likelihood that the learned content tag is relevant to the corresponding event type;
accessing a set of media having associated content tags and confidence scores;
determining that a subset of the plurality of event types are candidate event types for the set of media based at least in part on matching between the associated content tags of the set of the media and the set of learned content tags associated with each of the candidate event types;
generating, for a candidate event type of the candidate event types, a corresponding candidate event type score based at least in part on weighting the learned correlation score, for each learned content tag in the set of learned content tags for the candidate event type, by a corresponding one of the confidence scores for a corresponding one of the associated content tags of the set of media;
selecting an event type of the plurality of event types having a highest candidate event type score; and
causing a user interface to present a representation of the selected event type for the set of media.