本文算法利用亮度DCT DC值、宏块类型和运动向量来确定MPEG压缩视频中运动活动和视觉内容变化。
The proposed method determines the motion activity and visual content change of the MPEG compressed video by using the luminance DCT DC value, macroblock type and motion vector.
该算法只需要通过直接抽取MPEG视频流中的B帧、P帧中的宏块类型信息，并对其进行计算就可以检测出视频中存在的镜头边界，从而对场景变换进行精确定位。
The proposed method can detect scene boundaries through using various macroblock type of B-frame and P-frame in MPEG video flow in order to realize accurate scene change positioning.
应用推荐