Multi-speaker prosodic instance selection for HMM-based speech synthesis.
Yansuo YuFengyun ZhuXihong WuPublished in: ChinaSIP (2013)
Keyphrases
- speech synthesis
- prosodic features
- instance selection
- speech recognition
- hidden markov models
- text to speech
- vocal tract
- nearest neighbor
- feature and instance selection
- text classification
- multiple instance learning
- data reduction
- speaker verification
- semi supervised learning
- classification accuracy
- multi class
- automatic speech recognition
- supervised learning
- evaluation criteria
- knowledge discovery and data mining
- speech signal
- visual features
- pattern recognition
- data sets
- co training
- noisy environments
- pairwise
- language model
- image processing
- feature selection
- machine learning