Maximum a posteriori Speech Enhancement Based on Double Spectrum.
Pejman MowlaeeDaniel ScheranJohannes StahlSean U. N. WoodW. Bastiaan KleijnPublished in: INTERSPEECH (2019)
Keyphrases
- maximum a posteriori
- speech enhancement
- signal to noise ratio
- noise reduction
- noisy environments
- map estimation
- markov random field
- single channel
- image reconstruction
- maximum likelihood
- speech signal
- hyperparameters
- linear prediction
- edge preserving
- em algorithm
- multi channel
- vocal tract
- energy function
- additive noise
- wiener filter
- expectation maximization
- image processing
- parameter estimation
- higher order
- non stationary
- feature vectors