gmm ubm matlab,MSR Identity Toolbox工具箱中gmm_ubm_artificial的训练自定义声音来源

%{

This is a demo on how to use the Identity Toolbox for GMM-UBM based speaker

recognition. A small scale task has been designed using artificially

generated features for 20 speakers. Each speaker has 10 sessions

(channels) and each session is 1000 frames long (10 seconds assuming 10 ms

frame increments).

There are 4 steps involved:

1. training a UBM from background data

2. MAP adapting speaker models from the UBM using enrollment data

3. scoring verification trials

4. computing the performance measures (e.g., confusion matrix and EER)

Note: given the relatively small size of the task, we can load all the data

and models into memory. This, however, may not be practical for large scale

tasks (or on machines with a limited memory). In such cases, the parameters

should be saved to the disk.

Malcolm Slaney

Omid Sadjadi

Microsoft Research, Conversational Systems Research Center

%}

%

%%

% Step0: Set the parameters of the experiment

nSpeakers = 20;

nDims = 13;             % dimensionality of feature vectors

nMixtures = 32;         % How many mixtures used to generate data

nChannels = 10;         % Number of channels (sessions) per speaker

nFrames = 1000;         % Frames per speaker (10 seconds assuming 100 Hz)

nWorkers = 4;           % Number of parfor workers, if available

% Pick random centers for all the mixtures.

mixtureVariance = .10;

channelVariance = .05;

mixtureCenters = randn(nDims, nMixtures, nSpeakers);

channelCenters = randn(nDims, nMixtures, nSpeakers, nChannels)*.1;

trainSpeakerData = cell(nSpeakers, nChannels);

testSpeakerData = cell(nSpeakers, nChannels);

speakerID = zeros(nSpeakers, nChannels);

% Create the random data. Both training and testing data have the same

% layout.

for s=1:nSpeakers

trainSpeechData = zeros(nDims, nMixtures);

testSpeechData = zeros(nDims, nMixtures);

for c=1:nChannels

for m=1:nMixtures

% Create data from mixture m for speaker s

frameIndices = m:nMixtures:nFrames;

nMixFrames = length(frameIndices);

trainSpeechData(:,frameIndices) = ...

randn(nDims, nMixFrames)*sqrt(mixtureVariance) + ...

repmat(mixtureCenters(:,m,s),1,nMixFrames) + ...

repmat(channelCenters(:,m,s,c),1,nMixFrames);

testSpeechData(:,frameIndices) = ...

randn(nDims, nMixFrames)*sqrt(mixtureVariance) + ...

repmat(mixtureCenters(:,m,s),1,nMixFrames) + ...

repmat(channelCenters(:,m,s,c),1,nMixFrames);

end

trainSpeakerData{s, c} = trainSpeechData;

testSpeakerData{s, c} = testSpeechData;

speakerID(s,c) = s;                 % Keep track of who this is

end

end

%%

% Step1: Create the universal background model from all the training speaker data

nmix = nMixtures;           % In this case, we know the # of mixtures needed

final_niter = 10;

ds_factor = 1;

ubm = gmm_em(trainSpeakerData(:), nmix, final_niter, ds_factor, nWorkers);

%%

% Step2: Now adapt the UBM to each speaker to create GMM speaker model.

map_tau = 10.0;

config = 'mwv';

gmm = cell(nSpeakers, 1);

for s=1:nSpeakers

gmm{s} = mapAdapt(trainSpeakerData(s, :), ubm, map_tau, config);

end

%%

% Step3: Now calculate the score for each model versus each speaker's data.

% Generate a list that tests each model (first column) against all the

% testSpeakerData.

trials = zeros(nSpeakers*nChannels*nSpeakers, 2);

answers = zeros(nSpeakers*nChannels*nSpeakers, 1);

for ix = 1 : nSpeakers,

b = (ix-1)*nSpeakers*nChannels + 1;

e = b + nSpeakers*nChannels - 1;

trials(b:e, :)  = [ix * ones(nSpeakers*nChannels, 1), (1:nSpeakers*nChannels)'];

answers((ix-1)*nChannels+b : (ix-1)*nChannels+b+nChannels-1) = 1;

end

gmmScores = score_gmm_trials(gmm, reshape(testSpeakerData', nSpeakers*nChannels,1), trials, ubm);

%%

% Step4: Now compute the EER and plot the DET curve and confusion matrix

imagesc(reshape(gmmScores,nSpeakers*nChannels, nSpeakers))

title('Speaker Verification Likelihood (GMM Model)');

ylabel('Test # (Channel x Speaker)'); xlabel('Model #');

colorbar; drawnow; axis xy

figure

eer = compute_eer(gmmScores, answers, true);