void train (const ssi_char_t *dir, const ssi_char_t *model) {
// load samples
StringList files;
FileTools::ReadFilesFromDir (files, dir, "*.wav");
SampleList samples;
samples.addUserName ("user");
for (ssi_size_t i = 0; i < files.size (); i++) {
ssi_stream_t *stream = new ssi_stream_t;
ssi_sample_t *sample = new ssi_sample_t;
const ssi_char_t *filename = files.get (i);
// parse class name
FilePath fp (files.get(i));
ssi_char_t *class_name = ssi_strcpy (fp.getName ());
for (ssi_size_t j = 0; j < strlen (class_name); j++) {
if (class_name[j] == '_') {
class_name[j] = '\0';
break;
}
}
ssi_size_t class_id = samples.addClassName (class_name);
delete[] class_name;
// read wave file
WavTools::ReadWavFile (filename, *stream);
// create sample
sample->class_id = class_id;
sample->num = 1;
sample->score = 1.0f;
sample->streams = new ssi_stream_t *[1];
sample->streams[0] = stream;
sample->time = 0;
sample->user_id = 0;
// add sample
samples.addSample (sample);
}
// extract features
SampleList samples_t;
EmoVoiceFeat *ev_feat = ssi_create (EmoVoiceFeat, "ev_feat", true);
ModelTools::TransformSampleList (samples, samples_t, *ev_feat);
// create model
IModel *bayes = ssi_create (NaiveBayes, "bayes", true);
Trainer trainer (bayes);
// evalulation
Evaluation eval;
eval.evalKFold (&trainer, samples_t, 10);
eval.print ();
// train & save
trainer.train (samples_t);
trainer.save (model);
}
void ElanTools::LoadSampleList (SampleList &samples,
ssi_size_t num,
ssi_stream_t *streams[],
ElanTier &elanTier,
const ssi_char_t *user_name,
bool useTierNameAsLabel) {
// add user name
ssi_size_t user_id = samples.addUserName (user_name);
// add labels
ssi_size_t class_id;
if (useTierNameAsLabel) {
class_id = samples.addClassName (elanTier.name ());
}
// add samples
ElanTier::iterator anno;
for (anno = elanTier.begin (); anno != elanTier.end (); anno++) {
ssi_sample_t *sample = new ssi_sample_t;
ssi_stream_t **chops = new ssi_stream_t *[num];
bool success = false;
for (ssi_size_t j = 0; j < num; j++) {
// calculate start and stop index
ssi_size_t start_index = ssi_cast (ssi_size_t, (anno->from / 1000.0) * streams[j]->sr + 0.5);
ssi_size_t stop_index = ssi_cast (ssi_size_t, (anno->to / 1000.0) * streams[j]->sr + 0.5);
if (! (start_index <= stop_index && stop_index < streams[j]->num)) {
ssi_wrn ("invalid interval [%lf..%lf]s", anno->from/1000.0, anno->to/1000.0);
continue;
}
// extract sample
chops[j] = new ssi_stream_t;
ssi_stream_copy (*streams[j], *chops[j], start_index, stop_index);
success = true;
}
if (success) {
// create and add new sample
if (useTierNameAsLabel) {
sample->class_id = class_id;
} else {
sample->class_id = samples.addClassName (anno->value.str ());
}
sample->num = num;
sample->prob = 1.0f;
sample->streams = chops;
sample->time = anno->from / 1000.0;
sample->user_id = user_id;
samples.addSample (sample);
} else {
delete sample;
delete[] chops;
}
}
}
