Efficient Generation of high-order context-dependent Weighted Finite State Transducers for Speech Recognition.
Mike SchusterTakaaki HoriPublished in: ICASSP (1) (2005)
Keyphrases
- high order
- speech recognition
- context dependent
- higher order
- hidden markov models
- language model
- pairwise
- speech synthesis
- automatic speech recognition
- pattern recognition
- speech recognition technology
- low level
- noisy environments
- speech recognizer
- markov random field
- language modeling
- prediction accuracy
- natural language
- speaker identification
- speaker independent
- image processing
- feature vectors
- speech signal
- keywords
- image segmentation
- speaker adaptation