Research And Implementation Of Multi-Media Fusion Analyzing Technology

Description, management and manipulation of information content is the basic of information society, and there is a new challenge in the multimedia content management area. It is an important aspect of multimedia information management research to automatically analyze and extract the media content with computer. With vast multiple medias coming forth, the scheme to gain the content of multiple media by analyzing single media is deficiency for multimedia application, the way on multi-media fusion is inevitable. This paper focuses on the research of multimedia content analyzing with multi-media fusion approach. The main research content and contributions in this paper are as following: We systematically study the issue of multi-media fusion analyzing theory. The formal description of multi-media fusion analyzing approach is presented. We investigate the model and architecture of multi-media fusion analyzing system. We investigate content and structure analyzing of movie & TV programs employing multi-media fusion instrument in order to validate our discussion. Approaches of multi-media fusion scene detection and particular program content extracting are presented. The results and analysis from experience are given. We clear up the concept of film abstract. Investigate using multi-media fusion approach to resolve the problem of program content extraction and presentation. An approach of event-oriented feature abstracting is presented. We devise and improve some media processing algorithms including: text detection in video frames, audio shot detection, scene detection, evaluation of scene significance.Finally , the work background and application are introduced, the conclusions are made, and some promising research directions are given.