State-dependent mixture tying with variable codebook size for accented speech recognition.
Yi LiuFang ZhengLei HeYunqing XiaPublished in: ASRU (2007)
Keyphrases
- speech recognition
- state dependent
- steady state
- hidden markov models
- language model
- speech processing
- speech synthesis
- speech recognizer
- automatic speech recognition
- speech recognition technology
- pattern recognition
- vector quantization
- noisy environments
- speech signal
- queueing networks
- speech recognition systems
- optimal policy
- speaker independent
- computational complexity
- markov chain
- speech recognizers
- speaker dependent
- asymptotically optimal
- queue length
- stationary distribution
- speaker identification
- image compression
- arrival rate
- gaussian mixture model
- feature vectors
- mixture model
- learning algorithm
- audio visual speech recognition