Minimum Classification Error Interactive Training for Speaker Identification.
Yusuke KidaHiroyoshi YamamotoChiyomi MiyajimaKeiichi TokudaTadashi KitamuraPublished in: ICASSP (1) (2005)
Keyphrases
- minimum classification error
- speaker identification
- discriminative training
- speech recognition
- statistical pattern recognition
- gaussian mixture model
- hidden markov models
- noisy environments
- classification error
- maximum likelihood estimation
- speech signal
- gaussian mixture
- broadcast news
- training set
- feature extraction
- maximum likelihood
- generative model
- training samples
- probabilistic interpretation
- training examples
- data sets
- semi supervised
- multi class
- pattern recognition
- video sequences
- computer vision
- neural network