Generative Models for Improved Naturalness, Intelligibility, and Voicing of Whispered Speech.
Dominik WagnerSebastian P. BayerlHéctor A. Cordourier MaruriTobias BockletPublished in: SLT (2022)
Keyphrases
- markov random field
- generative model
- conditional random fields
- endpoint detection
- speech recognition
- probabilistic model
- higher order
- discriminative learning
- mixture model
- em algorithm
- discriminative models
- generative and discriminative models
- prior knowledge
- maximum entropy principle
- expectation maximization
- topic models
- object categories
- hierarchical hidden markov models
- speech signal
- hidden variables
- image processing
- weakly supervised
- semi supervised
- feature space
- computer vision
- data sets