Work

Analysis-by-synthesis frame dropping algorithm together with a novel speech recognizer using time-varying hidden Markov model