Speech Signals Identification base on Improved DBN

摘要：

　　For the problem low speech recognition rate,an improved method of combining Deep Belief Network(DBN)with support vector machine(SVM)for analyzing Small sample speech signals is proposed.The speech signal data collected as the training sample is used for training the DBN to get the optimal parameter values.The trained DBN is utilized for feature extraction,and these speech sample data signals will be classified using the SVM classification algorithm.The algorithm is tested through the simulation experiments under MATLAB.Experimental results indicate higher recognition rate and thus prove the effectiveness of the proposed DBN method in speech recognition field.

关键词： Speech recognition Deep Belief Network Support Vector Machine Features extraction

作者: CAI Jun YAO Qin ZHANG Yi

作者单位: Chongqing university of posts and telecommunications College of Automation,Chong qing 400065 Chongqing university of posts and telecommunications College of Advanced Manufacturing Engineering,C

会议类型: 国际会议

会议名称: 2017 IEEE 2nd Advanced Information Technology,Electronic and Automation Control Conference(IAEAC 2017)(2017 IEEE 第2届先进信息技术、电子与自动化控制国际会议)

会议地点: 重庆

会议语种:英文

页码: 1144-1148

在线出版日期: 2017-03-25（万方平台首次上网日期，不代表论文的发表时间）