Speech Recognition using FHMMS Robust Against Nonstationary Noise.
Agnieszka BetkowskaKoichi ShinodaSadaoki FuruiPublished in: ICASSP (4) (2007)
Keyphrases
- non stationary
- noisy environments
- speech signal
- white noise
- image noise
- random fields
- biomedical signals
- adaptive algorithms
- noise reduction
- stock price
- estimation error
- discrete valued
- autoregressive
- speech recognition
- missing data
- reinforcement learning
- concept drift
- noise level
- signal to noise ratio
- empirical mode decomposition
- scale space
- speech enhancement
- multiresolution
- pairwise