基于图像分割和对象跟踪的新闻视频镜头边界检测方法  

A Shot Boundary Detection Method for News Video Based on Image Segmentation and Object Tracking

作　　者：徐新文[1] 李国辉[1] 朱为[1] 

机构地区：[1]国防科技大学信息系统与管理学院,长沙410073

出　　处：《中国图象图形学报》2009年第8期1594-1600,共7页Journal of Image and Graphics

基　　金：国家自然科学基金项目(60243006);国家教育部博士点基金项目(20069998022)

摘　　要：镜头边界检测是许多多媒体应用的一个重要步骤,而现有的镜头边界检测方法大都是首先提取视频帧低层视觉特征,然后构造相异性测度函数,但由于这些方法对低层特征变化、对象运动、摄像机运动和视频质量较敏感,为克服此问题,提出了一种基于图像分割和对象跟踪的镜头边界检测方法。该方法首先采用分区直方图对镜头进行预检,然后利用基于小波分析的无监督图像分割和对象跟踪技术,通过构造相异性测度函数来对镜头边界进行确认。由于分区直方图方法作为第1过滤器,可有效地减少图像分割和对象跟踪的视频帧数目,从而提高了整个算法的效率,而基于小波变换的无监督图像分割和对象跟踪,则对以上问题具有较好的鲁棒性。在3个多小时的CCTV和CNN新闻视频实验中,获得了97.2%查准率和96.4%的查全率。As a critical step in many multimedia applications, shot boundary detection attracts many research interests in recent years. Most present methods measure the similarity among video frames based on its low-level feathers. However, they are sensitive to the change in brightness, color, motion of object,camera motions and the quality of video. This paper proposes an innovative shot boundary detection method for news video based on image segmentation and object tracking. It combines three main techniques, namely, the partitioned histogram comparison method, the image segmentation based on wavelet analysis and the object tracking. The partitioned histogram comparison is used as the first filter to effectively reduce the number of video frames which need segmentation and object tracking. The unsupervised image segmentation based on wavelet analysis and object tracking is robust to those problems mentioned above. The efficacy of the proposed method is extensively tested with more than 3 hours of CCTV and CNN news programs, and that 96. 4% recall with 97.2% precision has been achieved.

关 键 词：镜头边界检测 图像分割 对象跟踪 分区直方图 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN941.1[自动化与计算机技术—计算机科学与技术]