A discriminative performance metric for GMM-UBM speaker identification.
Omid DehzangiBin MaEngsiong ChngHaizhou LiPublished in: INTERSPEECH (2010)
Keyphrases
- speaker identification
- feature transformation
- gaussian mixture model
- feature extraction
- speech recognition
- speaker recognition
- probabilistic neural network
- speech signal
- noisy environments
- broadcast news
- mixture model
- audio features
- feature vectors
- em algorithm
- hidden markov models
- information retrieval
- expectation maximization
- maximum likelihood
- speaker diarization
- computer vision