摘要
新闻视频字幕是新闻视频数据中所包含的固有信息,是实现自动化视频检索的重要信息源,运用X^2直方图匹配法检测突变镜头,再通过C-均值聚类算法标识出新闻字幕,可以准确提取出新闻字幕,利用新闻视频字幕进行新闻视频检索,可以克服现有硬件设备的局限性,提高检索的速度和准确性,进而提高擞字电视的人机交互性。
News caption is inherent information which is contained in the news video data. It is one of the sources for the realization of automatic retrieval. Detecting abrupt scene by X2 histogram matching algorithm, and then marking news caption by C - means clustering algorithm can obtain news caption exactly. Using news caption for news video retrieval can overcome the localization of the hardware and improve the speed of retrieval, further more, improve alternation of digital television.
出处
《中国有线电视》
2006年第24期2414-2416,共3页
China Digital Cable TV
基金
广西自然科学基金项目(桂科自0135059)
广西科技攻关项目(桂科攻0322012-5)。
关键词
数字电视
新闻字幕
视频检索
digital television
news caption
video retrieval
