Context modeling with the stochastic segment model
Owen Kimball, Mari Ostendorf, Ibrahim Bechwati · IEEE Transactions on Signal Processing · 1992
An approach for context modeling in continuous speech recognition for models based on multivariate Gaussian distributions, specifically, the stochastic segment model, is described. Robust context models are obtained by typing distribution parameters across different classes of context. Experimental results in phoneme and word recognition are comparable to those achieved with discrete hidden Markov models using mixture distributions.>