摘要
针对传统镜头边界检测算法在阈值选取上的局限性,提出了基于镜头时间连续性的时间因子和基于镜头运动相关性的运动差异因子概念,并基于该双因子设定自适应阈值,将其运用于MPEG-2压缩域视频中.结果表明,该算法能够减少漏检和镜头运动造成的误检,有效提高镜头边界检测的准确率.
Traditional shot boundary detection methods share some limitation in threshold defining. The dual factors including time factor and motion difference factor were introduced. The dual-factor based adaptive threshold was put in the MPEG-2 video shot boundary detection method. The experiments verify that this method can reduce leak detections and false detections caused by camera motion effectively.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2009年第11期1685-1688,共4页
Journal of Shanghai Jiaotong University
基金
国家自然科学基金项目(60702042
60802057)
国家高技术研究发展计划(863)项目(2009AA01Z407)
关键词
时间因子
运动差异因子
自适应阈值
镜头边界检测
time factor
motion difference factor
adaptive threshold
shot boundary detection
