Speech feature compensation based on pseudo stereo codebooks for robust speech recognition in additive noise environments.
Tsung-hsueh HsiehJeih-Weih HungPublished in: INTERSPEECH (2007)
Keyphrases
- speech recognition
- additive noise
- speech signal
- noisy environments
- automatic speech recognition
- cepstral coefficients
- speech synthesis
- vocal tract
- hidden markov models
- speech processing
- speaker identification
- speech recognizer
- language model
- spatial domain
- speech enhancement
- linear prediction
- speech recognition systems
- pattern recognition
- vector quantization
- background noise
- speaker recognition
- speech recognition technology
- noisy images
- maximum likelihood
- speaker dependent
- noisy speech
- isolated word
- speaker independent
- random variables
- feature vectors
- speech recognizers
- acoustic features
- wiener filter
- point spread function
- motion estimation
- speech recognition errors
- face recognition
- image processing