Description: This is a speech recognition research tool. it
contains a feature extraction program and three speech
recognizers: a DTW recognizer, discrete didden Markov model (DHMM)
based recognizer and Continuous density hidden Markov mode (CHMM)
with Gaussian mixture functions based recognizer.
The utilities are grouped as:
feature -- extract featue vectors from a speech signal (MFCC etc.)
dtwcmp -- dynamic time-wapping (DTW) comparision.
gensym -- turn vector sequences to discrete observation symbols.
dhmm -- discrete HMM training program.
dtest -- DHMM companion test program.
chmm -- continuous density HMM training program.
viterbi -- CHMM companion test program.
Note, this is a research tool not a complete speech recognition system.