论文部分内容阅读
通常在我们描述一种关键帧提取方法时,总会有这样一种印象,即视频描述的运动强度越高,就越需要更多的关键帧才能表述其内容。实验证明,通过引用MPEG-7运动活力描述子和可靠性标准可以使用运动活力的强度值标示一个视频段。通过把该视频分镜分割成几个具有相同图像帧数的片断,然后对于每个片段取其位于中间的一帧图像为该段的关键帧,这样我们就可以得到整个视频分镜的关键帧。进一步的,我们可以根据经验分割出不同的片段数然后通过该方法,得到了需要的关键帧数并计算它们。相对于传统的基于颜色特征的关键帧提取算法,我们的这种方法要快速的多,因为它只是通过简单的计算和压缩区域提取就得到了关键帧。因此,确切来说它接近于理论上的最优方法。
Often when we describe a keyframe extraction method, there is always the impression that the higher the intensity of the video description, the more keyframes it needs to represent its content. Experiments have shown that by reference to MPEG-7 kinetic activity descriptors and reliability criteria, a video segment can be labeled using the intensity of the kinetic energy. By dividing the video mirror into several clips with the same number of frames, and then for each clip in the middle of a frame of the image as the key frame, so that we can get the entire video frame key frame . Further, we can split the number of different fragments based on experience and then through this method, we can get the required number of key frames and calculate them. Compared with the traditional color-based keyframe extraction algorithm, our method is much faster, because it only gets the keyframes through simple calculation and compression region extraction. Therefore, it is close to the theoretically optimal method.