Eigenvalues Driven Gaussian Selection in continuous speech recognition using HMMs with full covariance matrices.
Marko JanevDarko PekarNiksa JakovljevicVlado DelicPublished in: Appl. Intell. (2010)
Keyphrases
- covariance matrices
- covariance matrix
- maximum likelihood
- gaussian distribution
- hidden markov models
- gaussian mixture model
- vector space
- distance measure
- gaussian mixture
- feature vectors
- linear classifiers
- expectation maximization
- lie group
- multivariate normal
- principal component analysis
- objective function
- riemannian metric
- dimensionality reduction
- riemannian manifolds
- positive definite