Scaled factorial hidden Markov models: A new technique for compensating gain differences in model-based single channel speech separation.
Mohammad H. RadfarW. WongRichard M. DansereauWai-Yip ChanPublished in: ICASSP (2010)
Keyphrases
- hidden markov models
- single channel
- speech signal
- speech enhancement
- sound source
- speech recognition
- multi channel
- automatic speech recognition
- phoneme recognition
- noisy environments
- independent component analysis
- vocal tract
- speaker identification
- wavelet decomposition
- automatic speech recognition systems
- frequency domain
- audio visual
- speech synthesis
- linear prediction
- prior information
- noise reduction
- wiener filter
- sequential data
- visual speech
- signal processing
- audio features
- optical flow
- baum welch
- human computer interaction
- image compression
- visual information