HMM Parameter Adaptation Using the Truncated First-Order VTS and EM Algorithm for Robust Speech Recognition.
Haifeng ShenQunxia LiJun GuoGang LiuPublished in: CIS (1) (2005)
Keyphrases
- speech recognition
- em algorithm
- speaker adaptation
- expectation maximization
- hidden markov models
- maximum likelihood
- noisy environments
- mixture model
- gaussian mixture model
- parameter estimation
- speech signal
- maximum likelihood estimation
- speaker independent
- language model
- gaussian mixture
- keyword spotting
- speaker identification
- speech synthesis
- probabilistic principal component analysis
- automatic speech recognition
- speech recognizer
- expectation maximisation
- speech processing
- probabilistic model
- pattern recognition
- speech recognition technology
- maximum likelihood estimates
- generative model
- speech recognition systems
- density estimation
- maximum a posteriori
- probability density function
- speaker dependent
- image segmentation