Video Summarization using Singular Value Decomposition

In this work a novel video summarization system based on unique properties from singular value decomposition (SVD) is presented. Deriving compact representations of video sequences that are intuitive for users and let them easily and quickly browse large collections of video data is fast becoming one of the most important topics in content-based video processing. Such representations, are collectively refer to as video summaries. They rapidly provide the user with information about the contents of the particular video while preserving the information. The need for automatic methods for generating video summaries is fueled both from the user and production viewpoints. With the proliferation of personal video recorder devices and hand-held cameras, many users generate many times more video footage than they can digest. On the other hand, in today's fast-paced news coverage, programs such as sports and news must be processed quickly for production or their value quickly decreases. Such time constraints and the increasing number of services being offered place a large burden on production companies to process, edit, and distribute video material as quickly as possible.