pnnrengong.m

基于MATLAB实现的说话人识别程序

clear;
num=15;

disp('正在计算识别模型...')
for i=1:num
   fname = sprintf('%ss%d.wav', 'data\train\', i); 
   [s, fs] = wavread(fname);
    v =cfmfcc(s, fs); 
    [n,m]=size(v);
    a=reshape(v,1,n*m);
    b=a(1:700);
	ref(i).p = b;
    
end 
p=ref(1).p;
for i=1:num-1
    p=[p;ref(i+1).p];
end
T=[1 0 0 0 0 0 0 0 0 0 0 0 0 0 0;
   0 1 0 0 0 0 0 0 0 0 0 0 0 0 0;
   0 0 1 0 0 0 0 0 0 0 0 0 0 0 0;
   0 0 0 1 0 0 0 0 0 0 0 0 0 0 0;
   0 0 0 0 1 0 0 0 0 0 0 0 0 0 0;
   0 0 0 0 0 1 0 0 0 0 0 0 0 0 0;
   0 0 0 0 0 0 1 0 0 0 0 0 0 0 0;
   0 0 0 0 0 0 0 1 0 0 0 0 0 0 0;
   0 0 0 0 0 0 0 0 1 0 0 0 0 0 0;
   0 0 0 0 0 0 0 0 0 1 0 0 0 0 0;
   0 0 0 0 0 0 0 0 0 0 1 0 0 0 0;
   0 0 0 0 0 0 0 0 0 0 0 1 0 0 0;
   0 0 0 0 0 0 0 0 0 0 0 0 1 0 0;
   0 0 0 0 0 0 0 0 0 0 0 0 0 1 0;
   0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
];

p=p';
T=T';
spread=1;
net=newpnn(p,T,spread);


Y=sim(net,p);
yc=vec2ind(Y);

disp('正在计算测试识别的结果...')
for i=1:num
   fname = sprintf('%ss%d.wav', 'data\test\', i); 
   [s, fs] = wavread(fname);
    v =cfmfcc(s, fs); 
    [n,m]=size(v);
    a_test=reshape(v,1,n*m);
    b_test=a_test(1:700);
    test(i).p_test = b_test;
end 
p_test=test(1).p_test;
for i=1:num-1
    p_test=[p_test;test(i+1).p_test];
end
p_test=p_test';
Y_test=sim(net,p_test);
yc_test=vec2ind(Y_test);
yc
yc_test