Mandarin electrolaryngeal voice conversion with combination of Gaussian mixture model and non-negative matrix factorization.
Ming LiLuting WangZhicheng XuDanwei CaiPublished in: APSIPA (2017)
Keyphrases
- gaussian mixture model
- negative matrix factorization
- mixture model
- em algorithm
- matrix factorization
- speaker recognition
- principal component analysis
- sparse representation
- emotion recognition
- expectation maximization
- document clustering
- spectral clustering
- feature vectors
- maximum likelihood
- gaussian mixture
- feature space
- probability density function
- mouth region
- speaker identification
- speech recognition
- signal processing
- information retrieval systems
- probabilistic model
- image processing
- machine learning
- data sets