A new approach to speech enhancement by a microphone array using EM and mixture models.
Hagai AttiasLi DengPublished in: INTERSPEECH (2002)
Keyphrases
- mixture model
- speech enhancement
- expectation maximization
- em algorithm
- noise reduction
- noisy environments
- signal to noise ratio
- single channel
- gaussian mixture model
- generative model
- speech signal
- density estimation
- probabilistic model
- maximum likelihood
- vocal tract
- linear prediction
- unsupervised learning
- model selection
- mixture modeling
- finite mixtures
- background noise
- gaussian mixture
- wiener filter
- probability density function
- smoothing algorithm
- additive noise
- language model
- maximum a posteriori
- image coding
- speech synthesis
- graphical models
- image processing
- bayesian framework
- speech recognition
- bayesian information criterion
- edge detection
- reinforcement learning
- similarity measure