Speaker adaptation based on transfer vector field smoothing with continuous mixture density HMMs.

Kazumi Ohkura Masahide Sugiyama Shigeki Sagayama

Published in: ICSLP (1992)

Keyphrases

vector field
speech recognition
speaker adaptation
hidden markov models
optical flow
expectation maximization
speaker independent
gradient field
mixture model
critical points
velocity field
speech recognizer
speaker dependent
maximum likelihood
pattern recognition
gaussian mixture model
scalar field