两南大学硕十学位论文异,提取出运动信息,颜色,像素差等低层特征。然后将这些特征作为输入属性,将镜头所属的类型作为结果属性,,建立了低层特征到高层语义之间的规则,实现了良好的镜头分类效果。关键词::ZhangNanSupervisor:—qiangAbstractWiththedevelopmentofmultimediatechnologyandWebtechnology,moreandmoreinfomationisstored,,videocanbedividedintofourlevelsofvideo,scene,—-,shotboundaryisdetectedtosegmentvideointoshots;second,key仔锄esareextractedforthesimpleexpressionofcontentoftheshot;finally,thevideoinformationistagged,thatis,,dividingvideointoshots;anothersideisshotclassificationbasedcontent·Forshotboundarydetection,worksbasedapproachofshotboundarydetectionusingmulti·,basedonfeaturedifferencesbetweentwoadjacentframesandshiftingwindowrespectively,wereemployedtodetectabrupttransition,,works,,,,shotclassificationCanbedividedintotypeclassification,,wecla
视频镜头分割和其在视频检索中的应用 来自淘豆网m.daumloan.com转载请标明出处.
