Patent №
US 11,625,920
Granted
2023-04-11
Filed 2020
Owner
TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED
Lab
—
AI components
4
nlp · vision · kr · evo
Assignment
Recorded
Dataset
AIPD
2023_r1 edition
Application
16894086
Provided is a method for labeling a segment of a video, in a server. In the method, a multimedia file corresponding to an acting role is obtained. A role feature of the acting role is determined based on the multimedia file. A target video is decoded to obtain a data frame and a playing timestamp corresponding to the data frame, the data frame including at least one of a video frame and an audio frame. In the data frame of the target video, a target data frame that matches the role feature is identified. A segment related to performance of the acting role in the target video is automatically labeled based on a playing timestamp of the target data frame.
AI classification
Ownership
TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED
assignment · 528550319
Assignors
GAO, XIAOMENG
On an employer assignment, the assignors are typically the inventors.