摘要
提出一种基于隐马尔可夫模型(HMM)和学习向量量化(LVQ)神经网络的语音识别方法.该方法先用HMM生成最佳语音状态序列,然后用函数逼近技术产生对最佳状态序列进行时间归正,最后通过LVQ神经网络进行分类识别.理论和实验结果表明,混合模型的识别率明显高于隐马尔可夫模型的识别率.
Present s a new hybrid framework of hidden Markov models(HMM)and learning vector quantization(LVQ)neural networks for speech recognition.Here,the HMM is employed to produce a best speech state sequence which is warped to a fixed dimension vector and the LVQ neural network is used as classifier.The theoretical analysis and experimental results show that the new hybrid model leads to higher recognition rates than HMM.
出处
《微电子学与计算机》
CSCD
北大核心
2009年第3期49-51,共3页
Microelectronics & Computer
关键词
语音识别
隐马尔可夫模型
学习向量量化
混合模型
speech recognition
hidden Markov models
learning vector quantization
hybrid model
