Based Video Retrieval System With

With the rapid development of the multimedia technology and network technology, and the maturity of modern computer technology especially data storage and transmission technologies, the video as a indispensable information carrier is becoming more and more important in people's life , education , entertainment and so on. Although video has some advantages such as powerfully expression ability, large information in, vivid pictures and so on, but it also has some disadvantages such as non- structural data format, enormous data, opaque displaying of the content and more, those make the video analysis and management such as browsing and retrieving video. The Content-Based Video Retrieval technology can effectively organize and manage these data, store them to and get them from the storage in the multimedia data way, and make people find out the needed video segment from the big amounts video data easily. We know that video content includes the video structure information, the low-level audition and vision information, the high-level semantic information, all of these are the foundation of further editing and understanding. How to organize those integrated for a model that is used in video's analysis and retrieval, is still an extremely challenging subject in this domain at present.This text only discusses part of researching work we have done in the Content-Based Video Retrieval domain, the work mainly includes Content-Based Image Retrieval, Shot segmentation. Key frame extraction, and developed the simple system of content-based video retrieval.In this text, we first deal with the lower-level content information of video, extracted the color feature and texture feature from the key frame by using the Local Accumulation Histogram and building co-occurrence matrix to extract the contrast, angle second of matrix, entropy, correlation information, composed the color and texture feature to retrieval the Key frame, and then completed the Content-Based Image Retrieval;Afterward accomplished the organization of the video structure by the Shot segmentation which mainly include abrupt change which we used double windows-based algorithm and gradual transition which we used double threshold algorithm, and then we can finished Shotsegmentation through the coalition of them;then used the key frame to represent the segmented shot, and the video retrieval will transformed to the image retrieval;Finally, designed and developed the video retrieval system that based on the method we introduced above.