说话人识别系统(需要录音)(可以运行)
所属分类:数值算法/人工智能
开发工具:matlab
文件大小:2393KB
下载次数:15
上传日期:2019-05-22 13:11:41
上 传 者:
我爱吃肉条
说明: 这个代码是基于matlab开发的,能够实现语音说话人识别的代码
(This code is based on Python and can realize speech recognition.)
文件列表:
speaker_recognition\blockFrames.m (704, 2004-04-30)
speaker_recognition\demo1.m (3781, 2006-05-23)
speaker_recognition\disteu.m (771, 2004-04-16)
speaker_recognition\melfb.m (1333, 2004-04-16)
speaker_recognition\mfcc.m (1103, 2007-05-04)
speaker_recognition\test.m (1349, 2007-05-06)
speaker_recognition\train.m (1517, 2007-05-08)
speaker_recognition\vqlbg.m (1073, 2007-04-29)
speaker_recognition\mytest.m (479, 2007-05-08)
speaker_recognition\exp.txt (231, 2007-05-06)
speaker_recognition\vad.m (1728, 2007-05-05)
speaker_recognition\mytest.asv (478, 2007-05-08)
speaker_recognition\ENDPOINT.M (4301, 2007-05-04)
speaker_recognition\ABDTW.M (1242, 2004-11-06)
speaker_recognition\ABFEATUR.M (1219, 2004-11-06)
speaker_recognition\ABHMM.M (1319, 2004-11-06)
speaker_recognition\ABPREPRO.M (1225, 2004-11-06)
speaker_recognition\ABVQ.M (1341, 2004-11-06)
speaker_recognition\AVGMAG.M (567, 2004-11-06)
speaker_recognition\BACKWARD.M (872, 2004-11-06)
speaker_recognition\BACKWARD.MEX (8676, 2004-11-06)
speaker_recognition\BAUMWELC.MEX (9152, 2004-11-06)
speaker_recognition\BAUMWLCH.M (1532, 2004-11-06)
speaker_recognition\CEPCOEFF.M (766, 2004-11-06)
speaker_recognition\CEPFEA.M (906, 2004-11-06)
speaker_recognition\compareEndpoint.m (407, 2004-11-06)
speaker_recognition\comparefeaplot1.m (1453, 2004-11-06)
speaker_recognition\comparefeaplot2.m (1468, 2004-11-06)
speaker_recognition\comparefeaplot3.m (1452, 2004-11-06)
speaker_recognition\comparefeaplot3en.m (1223, 2004-11-06)
speaker_recognition\comparefeaplot.m (1454, 2004-11-06)
speaker_recognition\cwtem.m (452, 2004-11-06)
speaker_recognition\DELCEP.M (519, 2004-11-06)
speaker_recognition\diary (0, 2004-11-06)
speaker_recognition\DIST.M (214, 2004-11-06)
speaker_recognition\DSPLAB.BMP (32582, 2004-11-06)
speaker_recognition\DTW.M (2516, 2004-11-06)
speaker_recognition\DTWARP.M (720, 2004-11-06)
speaker_recognition\DTWold.MEX (11312, 2004-11-06)
... ...
在MATLAB环境下实现基于矢量量化的说话人识别系统。在实时录音的情况下,利用该说话人识别系统,对不同的人的1s~7s的语音进行辨识。实现与文本无关的自动说话人确认的实时识别。
使用说明:
1 训练
打开Matlab 使Current Directory为VQ所在的文件夹(比如:E:\vq)
在Command windows中输入
train('1s\',7)
这是将1s中的wav文件进行特征提取并产生VQ码本,
在workspace中有个ans的文件保存为.m文件,比如7.m,保存在1s文件夹下。(1s表示语音长1秒)
2 识别
在Command windows中输入
test('4s/',7,ans)
其中的ans 就是步骤1产生的码本。(必须保证7.m在workspace中打开)此时就会显示结果。
Speaker 1 matches with speaker 2
Speaker 2 matches with speaker 3
Speaker 3 matches with speaker 4
Speaker 4 matches with speaker 6
Speaker 5 matches with speaker 2
Speaker 6 matches with speaker 7
Speaker 7 matches with speaker 7
1s、4s分别是7个不同的人说话的录音,只是打乱了次序。
fid = fopen('code.bin','wb');
fwrite(fid,code,'integer*4');
近期下载者:
相关文件:
收藏者: