Subspace distribution clustering for continuous observation density hidden Markov models

This paper presents an eecient approximation of the Gaussian mixture state probability density functions of continuous observation density hidden Markov models (CHMM's). In CHMM's, the Gaussian mixtures carry a high computational cost, which amounts to a signii-cant fraction (e.g. 30% to 70%) of the total computation. To achieve higher computation and memory ee-ciency, we approximate the Gaussian mixtures by (a) decomposition into functions deened on subspaces of the feature space, and (b) clustering the resulting subspace pdf's. Intuitively, when clustering in a subspace of few dimensions, even few function codewords can provide a small distortion. Therefore, we obtain signiicant reduction of the total computation (up to a factor of two), and memory savings (up to a factor of twelve), without significant changes of the CHMMM's accuracy.