基于前向-后向HMM的连续语音识别系统的研究

Research on speech recognition system based on forward and backward HMM

在分析语音识别原理的基础上采用TMS320DM642 DPS芯片,利用前向-后向HMM声学模型和Viterbi算法进行模式训练和识别,设计了一种连续的、小词量的语音识别系统。实验结果表明,该语音识别系统具有较高的识别率和一定程度的鲁棒性,实验室识别率和室外识别率分别达到96.8%及91.2%,该语音识别系统具有良好的实用性和可移植性。

Based on the research for the principle of speech recognition and chips of TMS320DM642 DPS, using acoustic models of forward-backward HMM and Viterbi algorithm model to train model and recognition, then a speech recognition system of continuous and small vocabulary is established. Examples show that this voice recognition system has a high recognition rate and it is robust in a certain extent, the laboratory and outside recognition rate reache as high as 96.8% and 91.2%, this speech recognition system has a good practical or transplantation.