一种从压缩视频流中提取关键帧的方法

关键帧提取技术是视频分析和基于内容的视频检索的基础。关键帧的使用大大减少了视频索引的数据量,同时也为视频和检索提供了一个组织框架。论文首先简单介绍目前的关键帧提取技术,然后提出了一种基于DC系数和运动矢量直接从MPEG压缩视频提取关键帧的方法,无需全部解压,计算复杂度低,大大提高了提取速度。实验证明该方法计算量小,可以较好地代表视频内容。

An Approach of Key Frame Extraction from MPEG Compressed Video

Key frame extraction is one of fundamental processes in video analysis and content-based retrieval.Key frame can dramatically reduce the data of the video indexing,and provide a framework for video abstraction and retrieval.In this paper,the authors first briefly discuss the existing approaches,and present an approach on MPEG compressed streams to extract key frames with DCT DC coefficients and motion vectors.By performing minimal decoding on com-pressed bitstream,computational complexity is reduced,and the processing speed can be dramatically improved.Experi-mental results show that this proposed approach is computationally simple and able to adapt to the visual content.