| In recent years,with the continuous popularization of computers,computer data resources have been fully demonstrated and pictures,text messages and video messages can better convey human emotions.When users need video information for only a short time,they will spend unnecessary time and traffic getting the required video information from the linear video information.Therefore,it's been a hot topic that how to get the video information that users need from the linear video information quickly and effectively.This thesis focus on the video meta-information and use the video segmentation,video key frame extraction,video tagging technology to get the video directory information,thus solving the problem of non-linear video information browsing.First,to get the video meta-information processing framework,this thesis analyze the structure of the video data.Beginning with the acquisition of video meta-information,this thesis proposes the color information entropy algorithm based on the compression domain in the process of video segmentation and the information-based key frame extraction algorithm in the progress of the extraction of the key frame,thus completing the construction of video meta-information optimization framework.Second,to obtain the meta-information of video and carry on empirical analysis,this thesis chooses different kinds of video data.By comparing the results of applying the origin algorithm and the improved algorithm,the improved algorithm is proved to enhance the accuracy of the segmentation of video and the extraction of key frame.Finally,based on the results of the empirical analysis,this thesis presents the processing result of the optimized video meta-information processing framework in an XML file for use by the video directory system.Then,the design and implementation of video catalog system ViMeta is completed from the needs of users,the design of system platform,the detailed system design and the design of system interface to meet the demand of users to locate. |
PDF Full Text Request