Minimum classification error/eigenvoices training for speaker identification.
Fabio ValenteChristian WellekensPublished in: ICASSP (2) (2003)
Keyphrases
- minimum classification error
- speaker identification
- discriminative training
- statistical pattern recognition
- gaussian mixture model
- speech recognition
- classification error
- maximum likelihood estimation
- hidden markov models
- gaussian mixture
- noisy environments
- feature extraction
- speech signal
- broadcast news
- maximum likelihood
- training samples
- supervised learning
- neural network
- training set
- video sequences