void EM::read(const FileNode& fn)
{
Algorithm::read(fn);
decomposeCovs();
computeLogWeightDivDet();
}
void read(const FileNode& fn)
{
clear();
read_params(fn["training_params"]);
fn["weights"] >> weights;
fn["means"] >> means;
FileNode cfn = fn["covs"];
FileNodeIterator cfn_it = cfn.begin();
int i, n = (int)cfn.size();
covs.resize(n);
for( i = 0; i < n; i++, ++cfn_it )
(*cfn_it) >> covs[i];
decomposeCovs();
computeLogWeightDivDet();
}
void EM::eStep()
{
// Compute probs_ik from means_k, covs_k and weights_k.
trainProbs.create(trainSamples.rows, nclusters, CV_64FC1);
trainLabels.create(trainSamples.rows, 1, CV_32SC1);
trainLogLikelihoods.create(trainSamples.rows, 1, CV_64FC1);
computeLogWeightDivDet();
CV_DbgAssert(trainSamples.type() == CV_64FC1);
CV_DbgAssert(means.type() == CV_64FC1);
for(int sampleIndex = 0; sampleIndex < trainSamples.rows; sampleIndex++)
{
Mat sampleProbs = trainProbs.row(sampleIndex);
Vec2d res = computeProbabilities(trainSamples.row(sampleIndex), &sampleProbs);
trainLogLikelihoods.at<double>(sampleIndex) = res[0];
trainLabels.at<int>(sampleIndex) = static_cast<int>(res[1]);
}
}
