Show simple item record

Proc. Interspeech 2010

dc.contributor.authorLing, Zhen-Hua
dc.contributor.authorRichmond, Korin
dc.contributor.authorYamagishi, Junichi
dc.date.accessioned2010-12-22T11:14:26Z
dc.date.available2010-12-22T11:14:26Z
dc.date.issued2010en
dc.identifier.urihttp://hdl.handle.net/1842/4563
dc.description.abstractIn this paper we present a method to predict the movement of a speaker's mouth from text input using hidden Markov models (HMM). We have used a corpus of human articulatory movements, recorded by electromagnetic articulography (EMA), to train HMMs. To predict articulatory movements from text, a suitable model sequence is selected and the maximum-likelihood parameter generation (MLPG) algorithm is used to generate output articulatory trajectories. In our experiments, we find that fully context-dependent models outperform monophone and quinphone models, achieving an average root mean square (RMS) error of 1.945mm when state durations are predicted from text, and 0.872mm when natural state durations are used. Finally, we go on to analyze the prediction error for different EMA dimensions and phone types. We find a clear pattern emerges that the movements of so-called critical articulators can be predicted more accurately than the average performance.en
dc.titleHMM-based Text-to-Articulatory-Movement Prediction and Analysis of Critical Articulatorsen
dc.typeConference Paperen
rps.titleProc. Interspeech 2010en
dc.date.updated2010-12-22T11:14:26Z
dc.date.openingDate2010-09-22
dc.date.closingDate2010-09-24


Files in this item

This item appears in the following Collection(s)

Show simple item record