On the Impact of Annotation Errors on Unit-Selection Speech Synthesis.
Jindrich MatousekDaniel TihelkaLubos SmídlPublished in: TSD (2012)
Keyphrases
- speech synthesis
- speech recognition
- text to speech
- vocal tract
- semantic annotation
- active learning
- high impact
- prosodic features
- selection algorithm
- selection criteria
- real time
- genetic algorithm
- case study
- weakly labeled
- database
- automatic annotation
- automatic image annotation
- image acquisition
- image annotation
- bayesian networks
- metadata
- feature selection
- computer vision