A tree-based kernel selection approach to efficient Gaussian mixture model-universal background model based speaker identification.
Zhenyu XiongThomas Fang ZhengZhanjiang SongFrank K. SoongWenhu WuPublished in: Speech Commun. (2006)
Keyphrases
- gaussian mixture model
- speaker identification
- speaker recognition
- mixture model
- feature space
- feature vectors
- probabilistic neural network
- em algorithm
- speech recognition
- feature transformation
- maximum likelihood
- expectation maximization
- background subtraction
- feature extraction
- generative model
- speech signal
- background model
- noisy environments
- broadcast news
- principal component analysis
- support vector
- similarity measure