基于偶数帧段输入隐马尔可夫模型的噪声环境下汉语连续语音识别研究

Study on the Chinese continuous speech recognition under noise environments based on segmental unit input HMM

摘要: 提出了利用偶数帧段输入隐马尔可夫模型(HMM)提高在噪声环境下汉语连续语音识别系统鲁棒性的方法,并提出了对于传统谱相减降噪技术的修改方法。实验结果表明,本文的方法能有效地提高噪声背景下汉语连续语音识别系统的性能。

Abstract: This paper presents a method to improve the noise robustness of Chinese continuous speech recognition system based on Segmental unit input HMM, and to improve traditional spectral subtraction. Experimental results demonstrate the efficiency of the new algorithm.