Extracting demonstrations from in-situ video content

Abstract

Embodiments disclosed herein include a method, a non-transitory computer-readable medium, and a system for generating video clips for teaching how to apply a tools in various application programs for editing documents. The method includes identifying one or more characteristic features of a video clip. The method also includes providing the one or more characteristic features to a trained machine learning analysis module. The method further includes evaluating the characteristic features to generate a clip rating. The method also includes determining whether to discard the video clip based on the clip rating.