摘要
模型补偿技术已成功应用到噪声环境下的语音识别任务中。流行的模型补偿技术如Log-Add和Log-Normal PMC(并行模型合并)方法对动态特征参数通常只能给出近似的补偿。因此他们的识别率在较低的信噪比条件下变得很低。本文利用静态特征的导函数推导出了一种新的动态模型参数补偿方法。新的方法可以同任何已知的静态模型补偿算法结合产生出新的用于识别的噪声语音模型。实验证明这一新算法的应用,使其识别率比仅使用原有的模型补偿算法有较为明显的提高,并且新算法的复杂度较原有的模型补偿算法只有轻微的增加。
Model-based compensation techniques have been successfully used for speech recognition in noisy environments. Popular model-based compensation methods such as the Log-Add and Log-Normal PMC （Parallel Model Combination） generally use approximate compensation for dynamic parameters. Hence their recognition accuracy is degraded at low signal-to-noise ratios. A Dynamic Parameter Compensation Method （DPCM） is derived by means of the time derivatives of static features in this paper. The new compensated dynamic model together with any known compensated static model forms a new corrupted speech recognition model. Experimental results show that the recognition model using this DPCM scheme gives recognition accuracy better than the original model compensation method for different additive noises at the expense of slight increase in computational complexity.
出处
《电路与系统学报》
CSCD
北大核心
2008年第2期14-19,共6页
Journal of Circuits and Systems
基金
国家自然科学基金资助项目(60502041)
广东省自然科学基金博士启动资助课题项目(07300583)
关键词
鲁棒语音识别
模型补偿
动态参数补偿
robust speech recognition
model compensation
dynamic parameter compensation
