摘要
网络的普及和交互电视的应用推动了视频分类的发展,迫切需要一种方便、快速的自动视频分类方法。本研究利用从视频片段中提取的与镜头有关的特征、颜色特征、音频特征和运动特征作为视频内容分类的可计算特征,并基于粗糙集理论,发挥其无需先验信息而从信息系统中分析多余属性的能力和从决策表中抽取规则的能力,对上述可计算特征进行分类形成规则,从而实现对视频片段的分类。
The popularization of network and the application of interactive television have been promoting the development of video classification. A convenient and fast method of automatic video classification is needed. In our study, shot, color, audio and motion features of video clips were first extracted as the computable features for video classification. These features were then classified by making the use of the capabilities of analyzing redundant attributes and extracting rules from decision tables of rough set theory and the classification of video clips was finally achieved.
出处
《河北科技师范学院学报》
CAS
2009年第1期48-51,共4页
Journal of Hebei Normal University of Science & Technology
关键词
视频分类
特征提取
特征选择
粗糙集
video classification
feature extraction
feature selection
rough set
作者简介
曾晓宁（1973-），女，讲师。主要研究方向：多媒体技术及电子政务。
