Research On Key Frame Extraction In Content-based Video Retrieval

In the video information is highly inflated today, how to achieve rapid, effective,convenient query retrieval has become a highly popular research topic of video field.Video denote its characteristics with precise language, because it has a wealth ofinformation, so the artificial input video element attribute has a lot of shortcomings oftraditional. In this case, technology has become an urgent issue at present videocontent and video retrieval can be automatically extracted from the computer. Inrecent years, the rapid development of technology of content-based video retrieval,can satisfy the query requirements of the project, which is mainly dividedThe author’s main work is as follows:1. Several key technologies of content based video retrieval system is brieflyintroduced, including the analysis of video shot segmentation, structured, determinethe key frame and video feature extraction and matching algorithm. Then, introducethe video retrieval system similarity detection rules and video retrieval evaluationstandard.2. Based on the characteristics of the video, a video database model is complete.the proposed algorithm of adaptive lens luminance histogram is improved based onthe segmentation is analyzed and simulated, and the results of experimental resultsand current popular video segmentation algorithm are compared and analyzed,advantages it.3. Presents a key frame selection algorithm. The algorithm is different with thetraditional method is the key of the key frame selection algorithm based on the roughcontent variation degree of the lens to select1to3frames as the lens, the specificnumber of key frames automatically according to different degree of change isdecided to internal lens, thus greatly reducing the number of key frame redundancy.4. Presents an improved method of video retrieval, video retrieval methodscombined with the characteristics of the key frame and video sequence. The methodof better retrieval effect, a small amount of computation, the recall rate and accuracyrate is high, has the better anti disturbance.