连续语音识别网格技术在新闻制播平台的应用  

Continuous Speech Recognition Grid and Application of News Processing Platform

作　　者：张秋野[1] 王力劭[1] 丁鹏[2] 

机构地区：[1]中国国际广播电台,北京100040 [2]中国科学院自动化所,北京100080

出　　处：《电视技术》2010年第2期58-60,87,共4页Video Engineering

摘　　要：简要分析连续语音识别技术原理,介绍了语音识别网格构建海量多媒体新闻素材检索系统,该技术显著提升了多媒体新闻制播体系的素材资产化水平,为视音频媒体的多媒体内容资源检索带来了革命性变化。以中国国际广播电台(China Radio International,CRI)为例,描述了语音识别网格技术所带来的实际应用效果。Continuous speech recognition technology is analyzed, and the structure of the Continuous Speech Recognition Grid （CSRG） system for searching information in huge quantity of multimedia news materials is introduced. Technology in multimedia news processing platform is used, and it makes the materials be more valuable asset qualitatively. At last, taking the platform of CRI as an example, the actual result of the SRG is described.

关 键 词：连续语音识别 网格 多媒体 新闻 中国国际广播电台 

分 类 号：TN929.53[电子电信—通信与信息系统]