同时，MLLR可以通过大词汇连续语 音识别系统（Large vocabulary speech recognition system, LVCSR）进行聚类从而引入高层信息.
基于2个网页-相关网页
The system employs large vocabulary continuous speech recognition engine in front speech input end.
系统中的语音识别引擎使用的是大词汇量连续语音识别技术。
Large vocabulary continuous speech recognition system consists of several components, and recognition errors are caused by different factors.
大词表连续语音识别系统由多个组件构成，识别错误受多种因素的影响。
If speech sound of continuous large vocabulary can be automatically divided into segments of single syllable, the recognition ratio of system can be raised.
若能把连续大词汇量的语音进行实时自动切分为单个音节，便可提高系统的识别率。
应用推荐