ONLINE VIRTUAL COMMENTARY METHOD AND DEVICE, AND A MEDIUM

摘要

An online virtual commentary method and device, and a medium. The method comprises: acquiring, at least on the basis of a frame image of a video being played back, attribute data used for describing basic information of an element in the frame image (S101); extracting, on the basis of the attribute data, feature data used for representing comprehensive information in the frame image relevant to a commentary (S102); generating candidate events on the basis of the feature data (S103); selecting a commentary event from among the generated candidate events (S104); determining a corresponding commentary text on the basis of the selected commentary event (S105); and outputting corresponding commentary content on the basis of the commentary text (S106).