generic-library/opencv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Vadim, Maria, Alex, Andrey and I fixed the EM algorithm

Browse Source
This commit is contained in:
Ilya Lysenkov 2012-04-10 12:45:07 +00:00
parent b8f2011f80
commit ade6388d3a
1 changed files with 11 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
modules/ml/src/em.cpp
View File

@ -86,7 +86,8 @@ bool EM::train(InputArray samples,
OutputArray probs, OutputArray probs,
OutputArray logLikelihoods) OutputArray logLikelihoods)
{ {
setTrainData(START_AUTO_STEP, samples.getMat(), 0, 0, 0, 0); Mat samplesMat = samples.getMat();
setTrainData(START_AUTO_STEP, samplesMat, 0, 0, 0, 0);
return doTrain(START_AUTO_STEP, labels, probs, logLikelihoods); return doTrain(START_AUTO_STEP, labels, probs, logLikelihoods);
} }
@ -98,12 +99,13 @@ bool EM::trainE(InputArray samples,
OutputArray probs, OutputArray probs,
OutputArray logLikelihoods) OutputArray logLikelihoods)
{ {
Mat samplesMat = samples.getMat();
vector<Mat> covs0; vector<Mat> covs0;
_covs0.getMatVector(covs0); _covs0.getMatVector(covs0);
Mat means0 = _means0.getMat(), weights0 = _weights0.getMat(); Mat means0 = _means0.getMat(), weights0 = _weights0.getMat();
setTrainData(START_E_STEP, samples.getMat(), 0, !_means0.empty() ? &means0 : 0, setTrainData(START_E_STEP, samplesMat, 0, !_means0.empty() ? &means0 : 0,
!_covs0.empty() ? &covs0 : 0, _weights0.empty() ? &weights0 : 0); !_covs0.empty() ? &covs0 : 0, _weights0.empty() ? &weights0 : 0);
return doTrain(START_E_STEP, labels, probs, logLikelihoods); return doTrain(START_E_STEP, labels, probs, logLikelihoods);
} }
@ -114,9 +116,10 @@ bool EM::trainM(InputArray samples,
OutputArray probs, OutputArray probs,
OutputArray logLikelihoods) OutputArray logLikelihoods)
{ {
Mat samplesMat = samples.getMat();
Mat probs0 = _probs0.getMat(); Mat probs0 = _probs0.getMat();
setTrainData(START_M_STEP, samples.getMat(), !_probs0.empty() ? &probs0 : 0, 0, 0, 0); setTrainData(START_M_STEP, samplesMat, !_probs0.empty() ? &probs0 : 0, 0, 0, 0);
return doTrain(START_M_STEP, labels, probs, logLikelihoods); return doTrain(START_M_STEP, labels, probs, logLikelihoods);
} }
@ -337,7 +340,11 @@ void EM::clusterTrainSamples()
CV_Assert(meansFlt.type() == CV_32FC1); CV_Assert(meansFlt.type() == CV_32FC1);
if(trainSamples.type() != CV_64FC1) if(trainSamples.type() != CV_64FC1)
trainSamplesFlt.convertTo(trainSamples, CV_64FC1); {
Mat trainSamplesBuffer;
trainSamplesFlt.convertTo(trainSamplesBuffer, CV_64FC1);
trainSamples = trainSamplesBuffer;
}
meansFlt.convertTo(means, CV_64FC1); meansFlt.convertTo(means, CV_64FC1);
// Compute weights and covs // Compute weights and covs