节点文献
基于神经网络的矢量量化算法在语音辨识系统中的应用研究
The Research on Voice Recognition Based on Vector Quantization Combined with Neural Network
【摘要】 介绍了语音识别的一般概念,利用MFCC对语音信号的特征信息进行了提取,在分析VQ模型和人工神经元网络的基础上,提出了一种VQ模型与神经元网络相结合的方法,并且建立了相关的数学模型.通过MATLAB仿真证明该模型具有很好的稳定性.
【Abstract】 The conception of voice recognition is presented in this article.Based on the method of MFCC,the characteristics extraction of voice signal is introduced.Based on studying traditional VQ model and Neural network model,a way of using VQ combined with neural network is proposed,what is more a mathematical model is established.The experiment based on matlab show the the system possesses a stable quality.
【关键词】 语音识别;
特征提取;
神经网络;
MFCC;
【Key words】 voice recognition; characteristics extraction; neural network; MFCC;
【Key words】 voice recognition; characteristics extraction; neural network; MFCC;
【基金】 许昌学院青年预研基金资助项目(2007035)
- 【文献出处】 河南科学 ,Henan Science , 编辑部邮箱 ,2008年07期
- 【分类号】TN912.34
- 【被引频次】2
- 【下载频次】99