libTLK  1.3.1
Functions
Viterbi training from transcriptions of acoustic models
Algorithms

Functions

 tL_vtrainer_free (tLVTrainer *vtrainer)
 Frees memory.
 tL_vtrainer_new (const tLAModel *amodel, tLAMiCounts *counts, const tLSeqs *seqs, char **err)
 Gets a tLVTrainer.
 tL_vtrain (tLVTrainer *vtrainer, const tLFea *fea, const size_t seq)
 Updates acoustic model counts according to the Viterbi algorithm.

Function Documentation

tL_vtrain ( tLVTrainer *  vtrainer,
const tLFea fea,
const size_t  seq 
)

Updates acoustic model counts according to the Viterbi algorithm.

This function increases the acoustic model's counts, which are stored in the trainer, from one sample (feature vector sequence) and its transcription, using the Viterbi algorithm in order to obtain the integer counts. The sequence is specified as an integer code according to the sequence manager used during the creation of the viterbi trainer.

Parameters:
vtrainerThe viterbi trainer.
feaA feature vector manager, which must contain a sequence of feature vectors of an adequate type for the amodel.
seqThe code of the transcription of the sample.
Returns:
The sample probability for that transcription; -HUGE_VAL in case the sample cannot be used to update the counts.
tL_vtrainer_free ( tLVTrainer *  vtrainer)

Frees memory.

Frees the memory allocated for the auxiliary data.

Parameters:
vtrainerThe Viterbi trainer.
tL_vtrainer_new ( const tLAModel amodel,
tLAMiCounts counts,
const tLSeqs seqs,
char **  err 
)

Gets a tLVTrainer.

This function creates auxiliary data used to train an acoustic model from transcriptions using the Viterbi algorithm. This data can only be used to train the symbol sequences stored in seqs during the creation of the trainer. Currently supported types are: Bernoulli, DGaussian and Mixture.

Parameters:
amodelThe acoustic model.
countsThe integer counts which will be increased.
seqsA symbol sequence manager containing the symbol sequences which can be used to train samples.
errPointer to string variable. If not NULL, an error message is allocated in the variable in case of error.
Returns:
Viterbi trainer, or NULL in case of error.
 All Data Structures Variables