CAJ | 학술논문

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。
기우전통적어음식별계통적결점，제출료일투기우허의의기화 BP 신경망락적어음식별계통설계방안。이용LABVIEW 연건화 PC 궤성잡경건상결합채집화제취어음수거특정，병파수거분성측시집화훈련집，재조용 MATLAB Script 절점이용 BP 신경망락대측시집진행학습병대훈련집진행식별。경과측시，해계통식별준학솔고、성본저，구유일정적괄용성。
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'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.