探讨基于Tsallis熵的视频关键帧提取技术

The key frame extraction technology based on Tsallis entropy is discussed

关键帧是视频中的一组有限数量的帧的子集,一个视频的关键帧序列能够合理地概括该视频信息,从而减少过大的视频数据对生产生活带来的承载负重。本文讨论了基于Tsallis熵的Jensen距离公式——JTD在视频关键帧提取中的应用。根据得到的差异性距离值JTD,首先检查子镜头边界,进而从每个子镜头中抽取一帧作为该镜头的代表帧,最终得到该段视频的关键帧序列。

Key frame is a subset of a limited number of video frames and key frames of a video sequence can reasonably generalize the video information, thereby minimizing weight bearing on production and life bring large video data? This paper discusses Jensen distance formula based on Tsallis entropy of JTD application in video key frame extraction. According to the difference of the distance value JTD, first check the sub shot boundary, and then extracted from each sub shot frame as a representative of the lens frame, finally get the key frames of the video sequence.