%0 Electronic Article
%A J. Ming
%A F.J. Smith
%K speech recognition
%K joint probability distributions
%K spectral correlation
%K upper bound
%K stochastic correlation model
%X A stochastic model, drawn from the upper bound of the joint probability distributions, is suggested for modelling the spectral correlation in speech. Experiments on a speaker independent E-set database show the effectiveness of this new modelling approach.
%@ 0013-5194
%T Stochastic correlation model for speech recognition
%B Electronics Letters
%D May 1996
%V 32
%N 11
%P 970-971
%I Institution of Engineering and Technology
%U https://digital-library.theiet.org/;jsessionid=k2d9wx6rytsx.x-iet-live-01content/journals/10.1049/el_19960704
%G EN