基于数据挖掘的视频镜头分类方法  

Data Mining-based Video Shot Classification Method

作　　者：赵士伟[1] 卓力[1] 孙少卿[1] 沈兰荪[1] 

机构地区：[1]北京工业大学信号与信息处理研究室

出　　处：《北京工业大学学报》2012年第5期721-726,共6页Journal of Beijing University of Technology

基　　金：国家自然科学基金资助项目(61003289);北京市自然科学基金资助项目(4102008);教育部新世纪优秀人才支持计划资助项目(Q2002012200901)

摘　　要：提出了一种基于数据挖掘的视频镜头风格自动分类方法.该方法首先进行镜头边界检测和关键帧提取,然后基于关键帧和镜头分别提取了视频的颜色和运动等特征,并利用决策树技术在大量的训练数据中挖掘这些特征与镜头类别之间的潜在规律,最后利用这些规律对新的视频镜头进行分类.实验结果表明,与基于SVM的方法相比,本文方法不仅能获得较好的检测准确率,而且获取的规则易于理解.A novel classification method of video shot genre based on data-mining is proposed in this paper.First,shot boundary detection and key frames extraction are performed.Second,some visual features such as color and motion are extracted for the key frame and shots.Third,decision tree is applied to discover the rules between these features and shots classes from numerous training data.Finally,these rules are exploited to classify the new video shots.Experimental results show that compared with the method based on SVM（support vector machine）,the proposed method can achieve higher detection accuracy and the rules obtained are easy to comprehend.

关 键 词：视频镜头分类 镜头边界检测 关键帧提取 数据挖掘 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]