A fully-pipelined expectation-maximization engine for Gaussian Mixture Models.
Ce GuoHaohuan FuWayne LukPublished in: FPT (2012)
Keyphrases
- gaussian mixture model
- expectation maximization
- em algorithm
- mixture model
- maximum likelihood
- probabilistic model
- maximum likelihood estimation
- unsupervised learning
- density estimation
- gaussian mixture
- probability density function
- speaker recognition
- parameter estimation
- generative model
- k means
- gaussian distribution
- image segmentation
- maximum a posteriori
- feature vectors
- bayesian framework
- speaker identification
- variational bayes
- mixture of gaussians
- covariance matrices
- probability density
- feature space
- pattern recognition
- finite mixture model
- mixture modeling
- finite mixtures
- machine learning
- model based clustering
- classification accuracy
- log likelihood function
- minimum message length