一种基于运动特征的快速镜头边界检测方法

提出了一种基于MPEG视频流运动特征的镜头边界检测模糊推理方法。首先提取MPEG视频流中每帧的宏块信息，然后分析这些信息分别得到相似度、相似度差、运动活动性强度差、运动集中度差等，并将它们作为隶属度函数的输入量，按照一定的推理规则得到突变、渐变、无镜头变换三种情况的隶属度值，值最大者决定该帧是何种类型。实验结果表明该方法具有较高的检测精度，而且由于该方法不用对MPEG视频流进行解压缩，因此处理速度快，适合于实时应用场合。

Rapid motion feature based shot boundary detection method

An approach to shot boundary detection by fuzzy reasoning was introduced, which was based on the motion features of MPEG video sequence. Some information of macroblocks was extracted in each frame of MPEG video sequence. Then the information was analysed and the similarity, the difference of similarity, the difference of motion-activity intensity and the difference of the degree of motion concentration were got. Inputting these parameters and reasoning according to certain rules, the grade of membership of abrupt change, gradual change and no shot transition could be achieved. The result of experiments shows that this method has higher precision. Furthermore, when using this method, there's no need to decompress the compressed MPEG video sequence. So the speed of processing is very high, and it meets real-time applications.