Noise adaptive training for subspace Gaussian mixture models.
Liang LuArnab GhoshalSteve RenalsPublished in: INTERSPEECH (2013)
Keyphrases
- gaussian mixture model
- mixture model
- feature space
- covariance matrices
- speaker recognition
- feature vectors
- expectation maximization
- em algorithm
- noise level
- maximum likelihood
- variational bayes
- speaker identification
- probability density
- gaussian mixture
- training set
- high dimensional data
- low dimensional
- density estimation
- gaussian distribution
- data sets
- missing data
- signal to noise ratio
- principal component analysis
- high dimensional
- feature extraction
- supervised learning
- neural network