Acoustic Model Training Using Pseudo-Speaker Features Generated by MLLR Transformations for Robust Speaker-Independent Speech Recognition.
Arata ItohSunao HaraNorihide KitaokaKazuya TakedaPublished in: IEICE Trans. Inf. Syst. (2012)
Keyphrases
- speech recognition
- speaker independent
- speaker adaptation
- speech recognition systems
- speaker dependent
- acoustic models
- noisy environments
- speaker identification
- hidden markov models
- speech synthesis
- automatic speech recognition
- language model
- speech recognizer
- speech signal
- pattern recognition
- mel frequency cepstral coefficients
- maximum likelihood
- feature extraction
- generative model
- machine learning
- speaker recognition
- feature space
- low level
- speaker diarization
- digit recognition
- feature vectors
- information retrieval