摘要
镜头边界检测是进行视频内容分析的首要步骤 .提出一种新的基于分块向量角和模值二维直方图 (BAMH)的镜头边界检测方法 ,将视频帧分割为多块 ,分块统计向量角和模值二维直方图作为计算帧间不连续值的特征 ,采用自适应阈值分割 ,检测镜头边界 .实验结果表明 ,本文方法的检测性能优于全局向量角 -模值直方图方法 ,更优于
Dividing a video sequence into shots was the first step towards video content analysis and content based video browsing and retrieval. A novel method for shot boundary detection was presented based on block angle module histogram (BAMH). There, each video frame was divided into several blocks, angle module histograms are computed and normalized for blocks, followed by computing the distance between histograms of consecutive frames. Then an adaptive threshold was selected for detecting shot boundary. Experiments show that the performance of our method was higher than that of the HSV space histogram or globe angle module histogram.
出处
《小型微型计算机系统》
CSCD
北大核心
2004年第9期1652-1655,共4页
Journal of Chinese Computer Systems
基金
陕西省科技研究发展计划 ( 98k0 7-G4)资助
西安市科技攻关计划项目 ( GG990 7)资助
