Construction of News Video Summarization System by Conceptual Graph
Hidetoshi Hayashi, Ryong Lee, Yahiko Kambayashi · 2003
The high capacity of storage media and the development of compression technologies for video today have made it easy to store video data for many hours into second storage devices. The goal of this research is to summarize large quantity of news video to a kind of important topics after saving them. While there exist many researches relevant to video summarization, most of them focus mainly on technology of speech recognition or image analysis and compare video segments only by keywords. Therefore, the proposed news summarization system compares news articles based on formed not only by keywords but also by the connection between keywords. In this paper, we introduce an implemented system using CNN news as an example and show that Conceptual Graph does better in semantic summarization than keywords-only method. Keyword Video Summarization,Conceptual Graph,Text Mining