Speaker Identification Using Discriminative Learning of Large Margin GMM.
Khalid DaoudiReda JouraniRégine André-ObrechtDriss AboutajdinePublished in: ICONIP (2) (2011)
Keyphrases
- speaker identification
- discriminative learning
- gaussian mixture model
- generative model
- speech recognition
- mixture model
- speech signal
- feature extraction
- noisy environments
- structured prediction
- weakly supervised
- broadcast news
- maximum margin
- em algorithm
- probabilistic model
- image set
- object categories
- pairwise constraints
- feature vectors
- expectation maximization
- automatic speech recognition
- active learning
- noise reduction
- hidden markov models
- feature space
- audio features
- multi class
- image processing