Interpretable parametric voice conversion functions based on Gaussian mixture models and constrained transformations.
Daniel ErroAgustín AlonsoLuis SerranoEva NavasInma HernáezPublished in: Comput. Speech Lang. (2015)
Keyphrases
- gaussian mixture model
- mixture model
- speaker recognition
- maximum likelihood
- feature vectors
- expectation maximization
- gaussian mixture
- em algorithm
- probability density
- probability density function
- speaker identification
- variational bayes
- density estimation
- feature space
- non stationary
- generative model
- probabilistic model
- computer vision
- data mining