基于MATLAB和BP网络的语音识别系统

The speech recognition system based on MATLAB and BP neural networks

神经网络是近年来信息科学、脑科学、神经心理学等诸多学科共同关注和研究的热点。由于其具有良好的抽象分类特性,现已应用于语音识别系统的研究和开发,并成为解决识别相关问题的有效工具。文章在讲述语音识别过程的基础上重点讨论利用BP神经网络对语音进行识别,用MATLAB完成对神经网络的训练和测试,并获得满意的结果。

In recent years neural network is study hotspot of common attention on subject of information science, brain science, neu- ral psychology and so on. Because neural network has a good abstract classification characteristic, now it has been used in the re- search and development of speech recognition system, and becomes the valid tool which resolves a related problem gradually in speech recognition. This paper briefly introduces process of speech recognition and emphatically discusses how to recognize speech with BP Neural Networks, trains and tests the neural network in use of MATLAB, and acquires a satisfied result.