Vimeo, Inc.
Method and system for using semantic-segmentation for automatically generating effects and transitions in video productions

Abstract:

A method and a system for automatic video production are provided herein. The method may include the following steps: obtaining a set of media entities, wherein at least one of the media entities comprises a background and at least one foreground object; automatically analyzing the media entities using visual analysis; automatically selecting at least two visual portions, based on the visual analysis; computing, for at least one of the visual portions, semantic segmentation indicative of a support of the at least one foreground object, based on the visual analysis; generating at least one visual effect in which the foreground object and the background undergo two different visual operations; and generating a video production by combining a plurality of the visual portions into one video production, while including the at least one visual effect in the video production.