Cepstral Statistics Compensation Using Online Pseudo Stereo Codebooks for Robust Speech Recognition in Additive Noise Environments.
Jeih-weih HungPublished in: ICASSP (1) (2006)
Keyphrases
- speech recognition
- additive noise
- speech signal
- noisy environments
- hidden markov models
- automatic speech recognition
- language model
- noisy images
- linear prediction
- pattern recognition
- speech enhancement
- speech synthesis
- spatial domain
- background noise
- speech recognition technology
- speaker identification
- computer vision
- maximum likelihood
- speech recognizer
- vocal tract
- speech recognition systems
- random variables
- image patches
- vector quantization
- wiener filter
- denoising
- probabilistic model
- multiresolution
- feature selection