A statistical method for database reduction for embedded unit selection speech synthesis.
Pirros TsiakoulisAimilios ChalamandarisSotiris KarabetsosSpyros RaptisPublished in: ICASSP (2008)
Keyphrases
- database
- synthetic data
- high accuracy
- statistical model
- databases
- database systems
- pairwise
- cost function
- statistical information
- selection algorithm
- experimental evaluation
- dynamic programming
- significant improvement
- preprocessing
- computational complexity
- pattern recognition
- video sequences
- probabilistic model
- support vector machine
- mutual information
- statistical analysis
- support vector machine svm
- clustering method
- non stationary
- high precision
- similarity measure
- selection strategy
- neural network