INTERFRAME DEPENDENT HIDDEN MARKOV MODEL FOR SPEECH RECOGNITION

Authors
MING J SMITH FJ
Citation
J. Ming et Fj. Smith, INTERFRAME DEPENDENT HIDDEN MARKOV MODEL FOR SPEECH RECOGNITION, Electronics Letters, 30(3), 1994, pp. 188-189
Citations number
4
Categorie Soggetti
Engineering, Eletrical & Electronic
Journal title
Electronics Letters → ACNP
ISSN journal
00135194
Volume
30
Issue
3
Year of publication
1994
Pages
188 - 189
Database
ISI
SICI code
0013-5194(1994)30:3<188:IDHMMF>2.0.ZU;2-3
Abstract
A hidden Markov model (HMM) with first-order dependent observation den sities is presented to account for the statistical dependence between successive frames. A modified Viterbi algorithm is described to optimi se jointly the state sequence and dependence relation for the model pa rameter estimation as well as likelihood calculation. Preliminary expe riments show that this approach achieves better performance than the s tandard muitivariate Gaussian HMM.