摘要
镜头分割是基于内容的视频检索首先要解决的关键技术。一般聚类算法可能导致帧序不连续或分割错误,并且现有的分割算法基本上都是对镜头中连续帧依次进行比较,效率比较低;为了解决以上问题,提出了一种改进的镜头分割算法-基于聚类的间隔帧分割算法,在此算法引入了参考变量,镜头中的很多帧不需进行比较。实验表明本算法(采用颜色直方图作为切变判别依据)在镜头分割中提高了效率,并且提高了分割的准确率。
Shot segmentation is a vital technology that must be resolved firstly in video retrieval.Frame sequence＇s incontinuity or false segmentation can be caused by unsupervised clustering,and the existing techniques are based on sequential search,while is too expensive for practical use. To solve the problems above,an improved algorithm for shot segmentation based on clustering is proposed,a referenced variable is used in clustering to decide whether the shots continue clustering,and most video frames do not need to be compared.It can improve their performance substantially. Experimental results demonstrate that this method speeds up a conventional method based on color histograms on average while improving the accuracy of the segmentation.
出处
《微计算机信息》
2010年第3期6-7,10,共3页
Control & Automation
关键词
视频检索
镜头分割
聚类
直方图
video retrieval
shot segmentation
clustering
histogram
