基于虚拟仪器和BP神经网络的语音识别系统的设计

Design of speech recognition system based on virtual instrument and BP neural network

基于传统的语音识别系统的缺点,提出了一套基于虚拟仪器和BP神经网络的语音识别系统设计方案。利用LABVIEW软件和PC机声卡硬件相结合采集和提取语音数据特征,并把数据分成测试集和训练集,再调用MATLAB Script节点利用BP神经网络对测试集进行学习并对训练集进行识别。经过测试,该系统识别准确率高、成本低,具有一定的适用性。

Based on the shortcomings of the traditional speech recognition , this paper proposed a speech recognition system based on virtual instrument and the BP neural network .We used LABVIEW software and PC sound card hardware to collect and extract speech data＆#39;s characteristics, and divided the data into test and training sets , then using MATLAB Script node to do learning by means of BP neural network and to identify training sets on the test set .After testing, the system has accurate recognition rate and low cost , which has certain applicability.