Speech recognition using temporally connected kernels in mixture density hidden Markov models.
Panu SomervuoPublished in: ICASSP (2000)
Keyphrases
- hidden markov models
- speech recognition
- speech signal
- automatic speech recognition
- phoneme recognition
- latent variable models
- mixture distributions
- speaker independent
- automatic speech recognition systems
- text recognition
- keyword spotting
- speech processing
- audio signals
- kernel function
- sequential data
- viterbi algorithm
- visual speech
- conditional random fields
- fisher kernel
- speech recognizer
- random effects
- speaker identification
- markov models
- mixture model
- support vector
- acoustic models
- multiple kernel learning
- markov model
- gesture recognition
- feature space
- hidden state
- spatio temporal
- baum welch
- linear dynamical systems
- sequence classification
- visual speech recognition
- broadcast news
- hierarchical hidden markov model
- hidden states
- noisy environments