摘要： 提出一种基于粗糙集(RS)的压缩域关键帧提取算法。该算法从视频的压缩域数据流中提取I帧，以相邻I帧的差值为行、多个特征属性为列建立信息系统，对该系统进行归一化、离散化，采用RS的属性约简理论从中得到没有冗余的属性核，利用不可分辨关系进行关键帧提取。与像素差法、直流系数法的性能比较结果表明，该算法的计算复杂度更低，且适用于不同类型的视频。

关键词: 压缩域, 关键帧提取, 粗糙集, 信息系统, 不可分辨关系

Abstract: This paper proposes a key frame extraction algorithm based on Rough Set(RS) in compressed domain. It extracts I frame from compressed domain data sequence, and constructs information system with the difference between two adjacent I frames in column and attributes sets which are extracted from decompressed I frames in row, then the established information system is normalized and discredited. It simplifies information system by using attribute reduce theory of RS to obtain attribute cores without redundancy, and key frame is extracted by using the indiscernibility relation of RS. Experimental result shows that, competed with pixel difference algorithm and direct current coefficient algorithm, the algorithm proposed can reduce the complexity of computing and can be used for different types of video effectively.

Key words: compressed domain, key frame extraction, Rough Set(RS), information system, indiscernibility relation

中图分类号: 

• TP301.6