A method for automatically abstracting visual documents

摘要: Visual documents—motion sequences on film, videotape, and digital recordings—constitute a major source of information for the Space Agency, as well all other government private sector entities. This article describes method automatically selecting key frames from visual documents. These may in turn be used to represent total image sequence documents libraries, hypermedia systems, training guides. The performance abstracting algorithm reduces 51 minutes video 134 frames; reduction range 700:1. © 1993 John Wiley & Sons, Inc.