摘要
本文提出了基于 3维空间Viterbi算法的汉语连续语音识别方法 .本方法采用 6 0个音素单位的隐马尔可夫模型 (HMM)和 8个声调单位的HMM作为识别用基元模型 .音素基元模型和声调基元模型的识别结果的统合 ,采用音素单位的HMM状态、声调单位的HMM状态和时间的 3维空间Viterbi算法来实现 .语音声学处理和语音言语处理的结合 ,采用修改型Earley分析法的Top Done型文法分析器和OnePassDP为基础的帧同步识别算法来实现 .在由 10名话者发音的有关旅馆预约指南的识别困难度是 2 7 3的 10 70句子的识别实验中 ,总平均识别率达到 94 4% .
This paper presents a recognition method of Chinese continuous speech,in which p honetic and prosodic features are integrated in terms of 3-Dimension Viterbi se arch.The phonetic information is modeled by 60 phonemic HMMs and the prosodic in formation by 8 tone HMMs.Both recognitions are synchronized based on 3-Dimensio n Viterbi search.A frame-synchronous parsing algorithm for CFG based on a top- down strategy is used for parsing processing.The task is related to the hotel re servation process,of which the perplexity is 27 3.For 1070 utterances produced by each of ten speakers,the average sentence recognition rate was 94 4%.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2000年第7期67-69,58,共4页
Acta Electronica Sinica
基金
国家自然科学基金资助课题
关键词
汉语连续语音识别
三维空间
VITERBI算法
recognition of Chinese continuous speech
phonemic HMM
tone HMM
3-Dimension Vite rbi search
