Maximum likelihood PSD estimation for speech enhancement in reverberant and noisy conditions.
Adam KuklasinskiSimon DocloJesper JensenPublished in: ICASSP (2016)
Keyphrases
- maximum likelihood
- speech enhancement
- noisy environments
- speech signal
- parameter estimation
- additive noise
- speech recognition
- noise reduction
- automatic speech recognition
- em algorithm
- expectation maximization
- linear prediction
- maximum likelihood estimator
- vocal tract
- background noise
- signal to noise ratio
- maximum a posteriori
- frequency domain
- hidden markov models
- spatial domain
- non stationary
- wiener filter
- language model