Study On IPTV Streaming Media Content - Based Retrieval Prototype

With the development of multimedia technology, building a retrieval system which can retrieve great capacity video information fast and efficiently becomes an immediate issue to resolve in information ages. However, because of the various varieties in multimedia format, the video retrieval system standardized function depresses. So the MPEG has presented a new standard MPEG-7 for various multimedia information descriptions to realize the multimedia information retrieval rapidly and effectively.Presently,a new technology ,the Internet Protocol Television(IPTV) makes use of the streaming media technology, transmits digital TV signal to users by Wide Band and supplies many alternating services, has more attentions in IT industry. In the IPTV system, when people demand program on TV, they need to know not only the program name, but also know the approximate scenario of program, or choose the actor or style by the favor of themselves. So it needs content-based video retrieval, which extracts given information clue from media data, then searches from the multimedia database and at last retrieval similar feature media data.Based on above requirements, this thesis makes research on them correspondingly, and then presents a video retrieval system model based on MPEG-7 which applies in IPTV system.It sets up an MPEG-7-based video description hiberarchy. Then in the system shot detection, it integrates many temporal segmentation algorithms to detect the shot. At last, it improves the IPTV system, and adds the content-based video retrieval system.The model constructed by this thesis based on MPEG-7, has standardized capability, at the same time can deal with different media formats. And its search capability has rather high level in precision and recall.