Automatic video generation using auto-adaptive video story models

摘要

A video processing system generates and automatically updates videos. A client inputs a smart script that defines general preferences and information to include in a video. The video processing system generates a model of the story from the smart script. A story model comprises story beats that define order and characteristics of information that is presented in a story. The video processing system accesses client content (e.g., video clips, images, and testimonials from a client website) and assigns content items to story beats. The story model and associated content is rendered into a video. The video processing system can then adapt the story model in response to viewer feedback, external events, or other client-supplied parameters. Adapting the story model may include reordering, adding, or removing story beats, or altering parameter values associated with story beat characteristics. Thus, the video processing system can iteratively improve and update a video automatically.