Modelling non-stationary noise with spectral factorisation in automatic speech recognition.
Antti HurmalainenJort F. GemmekeTuomas VirtanenPublished in: Comput. Speech Lang. (2013)
Keyphrases
- non stationary
- speech signal
- spectral subtraction
- automatic speech recognition
- noisy environments
- linear prediction
- speech recognition
- vocal tract
- word error rate
- broadcast news
- spoken words
- speech corpus
- autoregressive
- adaptive algorithms
- speaker adaptation
- speech retrieval
- conversational speech
- white noise
- hidden markov models
- recognition errors
- empirical mode decomposition
- spontaneous speech
- word recognition
- image compression
- biomedical signals
- speech sounds
- data mining
- noise model
- noise level
- noise reduction
- computer vision