Co-channel speaker separation based on maximum-likelihood deconvolution.
Michael I. SavicHuiqin GaoJeffrey S. SorensenPublished in: ICASSP (1) (1994)
Keyphrases
- maximum likelihood
- speaker adaptation
- multi channel
- least squares
- em algorithm
- speaker recognition
- speech recognition
- denoising
- speaker verification
- expectation maximization
- likelihood function
- audio visual
- gaussian distribution
- image deconvolution
- gaussian mixture model
- image restoration
- estimation algorithm
- additive noise
- speaker diarization
- blind deconvolution
- motion deblurring
- channel coding
- speaker identification
- feature vectors
- fir filters
- ofdm system
- inverse problems
- maximum a posteriori
- blurred images
- motion blur