Video Browsing And Retrieval Of Key Technologies Of Media Asset Management System

Media asset management is a total solution to managing all kinds of media such as video and audio data, text, files, figures and tables. It can satisfy the requirement of media asset owner to collect, keep, search ,edit and issue various information and it can also provide a simple method of visiting the contents for the media asset user in order to keep and use media asset high efficiently. It is a new concept in technique field of broadcast and TV system, at the same time, it is also the necessity of broadcast and TV digitalizing. The needs and requirements of Video Browsing and Retrieval are going on upgrading after the exploding of multimedia amounts, especially for the video media. How to find the content that oneself needs fast, effectively, have already become the urgent demand. This thesis deals with shot segmentation and key frame extraction, which are two main tasks of Video Browsing and Retrieval.Based on the analysis of shot segmentation, we proposes a new algorithm, which is based on the combination of Histogram, Gray difference frames and Automatic Threshold Detection. Experiments show that this method can train out specific model for different type of videos, while the speed needs to be enhanced. And it can express the main content of the video better. And then, we described the normal questions of the Video Similarity Calculation. Considering the user-aware and the shot transform, we combines some characteristics in the video. We proposes a new video similarity calculation method which can improve the retrieval precision. At last, we do some discussion in key frame extraction and introduce the functions of the Media asset management and the subsystems.