Two-stage phone duration modelling with feature construction and feature vector extension for the needs of speech synthesis.
Alexandros LazaridisTodor GanchevIosif MporasEvangelos DermatasNikos FakotakisPublished in: Comput. Speech Lang. (2012)
Keyphrases
- feature construction
- speech synthesis
- feature vectors
- speech recognition
- text to speech
- concept learning
- genetic programming
- data representation
- decision trees
- rule learning
- machine learning
- supervised learning
- feature space
- feature extraction
- feature selection
- dimension reduction
- support vector machine
- image processing
- data mining
- language model
- domain knowledge
- support vector
- information retrieval