摘要
计算机唇读是利用计算机对说话者的唇动等视觉语音信息进行分析以识别出其所说内容的过程,并可与听觉语音信息相融合以进一步提高计算机的识别率,从而使人机交互更加自然。本文从计算机唇读系统的各环节入手综述了该领域的研究进展,并讨论了现有诸方法的优缺点,最后提出了有待进一步研究的问题。
As a hotspot in the field of human-computer interaction, computer lip-reading aims at recognizing what human says by analyzing visual speech information, such as lip movement. It can be further integrated with audio speech information to improve recognition accuracy for more convenient human-computer interaction. This paper gives a survey of lip-reading approaches and discusses their benefits and drawbacks. Finally, several key issues to be researched in the field are pointed out.
出处
《数据采集与处理》
CSCD
北大核心
2007年第3期353-359,共7页
Journal of Data Acquisition and Processing
基金
国家自然科学基金(60121101)资助项目
关键词
唇读
定位
特征抽取
信息融合
lip-reading
location
feature extraction
information fusion