Exemplar-based speech enhancement for deep neural network based automatic speech recognition.
Deepak BabyJort F. GemmekeTuomas VirtanenHugo Van hammePublished in: ICASSP (2015)
Keyphrases
- speech enhancement
- automatic speech recognition
- speech signal
- noisy environments
- speech recognition
- vocal tract
- linear prediction
- neural network
- word error rate
- background noise
- hidden markov models
- speaker identification
- conversational speech
- additive noise
- spectral subtraction
- broadcast news
- speaker verification
- acoustic features
- non stationary
- speech retrieval
- noise reduction
- signal to noise ratio
- sound source
- speech synthesis
- higher order
- high resolution
- feature extraction