Do GMM Phoneme Classifiers Perceive Synthetic Sibilants as Humans Do?
Gábor PintérHiroki WatanabePublished in: INTERSPEECH (2016)
Keyphrases
- gaussian mixture model
- support vector
- training data
- speaker dependent
- speaker identification
- feature selection
- multiple classifiers
- test set
- supervised classification
- real world
- classification systems
- decision trees
- naive bayes
- speech recognition
- human subjects
- classification models
- svm classifier
- human users
- human behavior
- em algorithm
- feature vectors
- classification algorithm
- ensemble learning
- class labels
- feature set
- classifier ensemble
- hidden markov models
- training set
- gaussian mixture modeling