Based On The Content Of News Video Retrieval System Technology Research

News video contains lots of useful information, so how to classify, organize and retrieve news video has become an important research field. Due to-the large amount of data involved, it's quite time-consuming and subjective-judgement sensitive to classify news video with manual work. Therefore, developing an automatic analyzing method has become a significant research subject. The content of this paper mainly consists of two parts: news video retrieval and news video query. As the base of video query, news video retrieval is divided into structurized analysis and semantic exaction of video information. The former is composed of news item segmentation, frame segmentation and key frame extraction and the latter is composed of TV-station label recognition, caption extraction, audio analysis and visual semantic exaction. This paper comprises the basic functions that a news video retrieval system based on semantic should have, which provides a good technological foundation for further research on news video retrieval.