A maximum likelihood neural network based on a log-linearized Gaussian mixture model.
Toshio TsujiHiroyuki IchinobeOsamu FukudaMakoto KanekoPublished in: ICNN (1995)
Keyphrases
- gaussian mixture model
- maximum likelihood
- em algorithm
- mixture model
- expectation maximization
- speaker recognition
- gaussian distribution
- likelihood function
- speaker identification
- maximum likelihood estimation
- maximum a posteriori
- background subtraction
- gaussian mixture
- mixture distribution
- variational bayes
- density estimation
- gaussian mixture modeling
- feature vectors
- probabilistic model
- active learning