On-Line Video Summarization Based on Signature-Based Junk and Redundancy Filtering

Abstract

The huge amount of video content and the increasing rate in which new content is being created makes video summarization a necessary tool for the efficient access and management of video repositories. Nevertheless, the huge amount of data and the complexity of current techniques make the implantation of the summarization systems a difficult or non practical task. This paper presents current advances in the development of a novel algorithm for on-line video summarization, which can be easily used also as a video retrieval tool, aimed to combine summarization quality with high computational efficiency providing a system that can be practical for both commercial and home environments.