Improving speech recognition performance via phone-dependent VQ codebooks and adaptive language models in SPHINX-II.
Mei-Yuh HwangRonald RosenfeldEric H. ThayerMosur RavishankarLin Lawrence ChaseRobert WeideXuedong HuangFil AllevaPublished in: ICASSP (1) (1994)
Keyphrases
- speech recognition
- language model
- vector quantization
- vector quantizer
- language modeling
- probabilistic model
- acoustic models
- image compression
- retrieval model
- document retrieval
- automatic speech recognition
- speech recognizer
- n gram
- test collection
- information retrieval
- query expansion
- speech signal
- handwriting recognition
- speaker recognition
- word error rate
- speech recognition systems
- spoken term detection
- noisy environments
- hidden markov models
- image coding
- mixture model
- relevance model
- speaker identification
- translation model
- speaker independent
- visual words
- wavelet transform
- pattern recognition
- neural network