新闻视频中广告片段精确定位方法研究  

Study on Precise Commercial Segments Location in News Video

作　　者：李士进[1] 郭跃飞[2] 李昊[1] 秦明举[1] 冯钧[1] 

机构地区：[1]河海大学计算机及信息工程学院,南京210098 [2]复旦大学计算机科学与工程系,上海200433

出　　处：《中国图象图形学报》2009年第7期1432-1439,共8页Journal of Image and Graphics

基　　金：国家自然科学基金项目(60673141)

摘　　要：由于广告节目制作风格的多样性,新闻视频中广告片段的检测和定位是一个非常具有挑战性的问题。提出了一种新闻视频中广告片段定位的新方法。首先针对镜头切换和非镜头切换数目不平衡问题,设计了一种新的分类方法进行镜头检测,然后利用聚类分析粗略地标识广告块,最后通过分析相邻镜头的平均持续时间和镜头关键帧的视觉特征准确定位广告边界。实验结果表明,该方法具有较高的定位精度。The detection and location of commercials in news video has been a challenging problem because of the diversity of the production styles of commercial programs. In this paper, the authors present a novel algorithm for the location of commercial segments in news program. With the proposed method, firstly shot transition detection is conducted, which is based on the imbalanced data classification methodology, and then clustering analysis is employed to label commercial blocks roughly. Finally the accurate boundaries between the commercial segments and news story are located by analyzing a new set of features, including the average duration of preceding and subsequent shots and the visual features of the shots, such as color, saturation and edge distribution. The experiment results show that the proposed algorithm is effective with high precision.

关 键 词：广告片段定位 聚类 时间和视觉特征 镜头检测 不平衡数据 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]