Parametric subspace modeling of speech transitions

This paper describes an attempt at capturing segmental transition informati on for speech recognition tasks. The slowly varying dynamics of spectral tr ajectories carries much discriminant information that is very crudely model led by traditional approaches such as HMMs. In approaches such as recurrent neural networks there is the hope, but not the convincing demonstration, t hat such transitional information could be captured. The method presented h ere starts from the very different position of explicitly capturing the tra jectory of short time spectral parameter vectors on a subspace in which the temporal sequence information is preserved. This was approached by introdu cing a temporal constraint into the well known technique of Principal Compo nent Analysis (PCA). On this subspace, an attempt of parametric modelling t he trajectory was made, and a distance metric was computed to perform class ification of diphones. Using the Principal Curves method of Hastie and Stue tzle and the Generative Topographic map (GTM) technique of Bishop, Svensen and Williams as description of the temporal evolution in terms of latent va riables was performed. On the difficult problem of /bee/, /dee/, /gee/ it w as possible to retain discriminatory information with a small number of par ameters. Experimental illustrations present results on ISOLET and TIMIT dat abase. (C) 1999 Elsevier Science B.V. All rights reserved.