Objective evaluation of HMM-based speech synthesis system using kullback-leibler divergence.
Cong-Thanh DoMarc EvrardA. LemanChristophe d'AlessandroAlbert RilliardJ.-L. CrebouwPublished in: INTERSPEECH (2014)
Keyphrases
- objective evaluation
- kullback leibler divergence
- speech synthesis
- hidden markov models
- speech recognition
- ground truth
- subjective evaluation
- mutual information
- text to speech
- probability density function
- information theoretic
- information theory
- distance measure
- marginal distributions
- diffusion tensor
- similarity measure
- image segmentation
- conditional random fields