SABR: sparse, anchor-based representation of the speech signal.
Christopher LiberatoreSandesh AryalZelun WangSeth PolsleyRicardo Gutierrez-OsunaPublished in: INTERSPEECH (2015)
Keyphrases
- speech signal
- speech recognition
- machine learning
- non stationary
- speaker identification
- hidden markov models
- gaussian mixture model
- automatic speech recognition
- noisy environments
- spectral analysis
- linear prediction
- language model
- blind source separation
- background noise
- vocal tract
- automatic speech recognition systems