Unsupervised prosody labeling for constructing Mandarin TTS.
Chen-Yu ChiangSin-Horng ChenYih-Ru WangPublished in: SSW (2010)
Keyphrases
- text to speech
- prosodic features
- speech synthesis
- unsupervised learning
- speaker verification
- speech recognition
- active learning
- supervised learning
- probabilistic relaxation
- word processing
- unsupervised manner
- real time
- image labeling
- supervised classification
- data driven
- multiresolution
- data sets
- spontaneous speech
- database
- semi supervised learning
- multi modal
- labeling scheme
- pattern recognition
- image segmentation
- data mining
- information systems
- spoken document retrieval
- search engine
- machine learning