A fast, accurate approximation to log likelihood of Gaussian mixture models.
Pierre L. DogninVaibhava GoelJohn R. HersheyPeder A. OlsenPublished in: ICASSP (2009)
Keyphrases
- log likelihood
- gaussian mixture model
- em algorithm
- density estimation
- maximum likelihood
- mixture model
- expectation maximization
- variational bayes
- probability density function
- speaker recognition
- closed form
- information theoretic
- parameter estimation
- scoring function
- exponential family
- maximum likelihood estimation
- feature space
- feature vectors
- probabilistic model
- computer vision
- maximum a posteriori
- high dimensional
- density function
- pairwise
- speaker identification
- feature extraction