摘要
提出了一种针对MPEG视频的精确定位镜头边缘的方法,结合视频非压缩域和压缩域中相关视频帧的信息来定位镜头边缘。首先利用解压后的I帧信息来粗略定位镜头边缘,然后利用视频序列压缩域中的B帧的运动趋势来精确定位镜头边缘,而不用对其解压缩。实验结果表明此方法能精确地检测到视频的镜头边缘。
Shot boundary detection is the first step of video retrieval. A method of accurate shot boundary detection is proposed in this paper. First shot boundary is generally located by I frame , then it is accurately detected by the two sequential B frames. Experimental resuits prove that it is an efficient algorithm in shot boundary detection.
出处
《计算机技术与发展》
2007年第7期102-104,107,共4页
Computer Technology and Development
关键词
视频检索
镜头边缘检测
压缩域
video retrieval
shot boundary detection
compressed domain
