Acoustic-to-Articulatory Inversion based on Local Regression
2010 (English)In: Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010, Makuhari, Japan, 2010, 937-940 p.Conference paper (Refereed)
This paper presents an Acoustic-to-Articulatory inversionmethod based on local regression. Two types of local regression,a non-parametric and a local linear regression have beenapplied on a corpus containing simultaneous recordings of positionsof articulators and the corresponding acoustics. A maximumlikelihood trajectory smoothing using the estimated dynamicsof the articulators is also applied on the regression estimates.The average root mean square error in estimating articulatorypositions, given the acoustics, is 1.56 mm for the nonparametricregression and 1.52 mm for the local linear regression.The local linear regression is found to perform significantlybetter than regression using Gaussian Mixture Modelsusing the same acoustic and articulatory features.
Place, publisher, year, edition, pages
Makuhari, Japan, 2010. 937-940 p.
Acoustic-to-articulatory inversion, K-nearest neighbours, Local regression
Computer Science Language Technology (Computational Linguistics)
IdentifiersURN: urn:nbn:se:kth:diva-52155ISI: 000294382400231ScopusID: 2-s2.0-79959816993ISBN: 978-1-61782-123-3OAI: oai:DiVA.org:kth-52155DiVA: diva2:465450
11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010, Makuhari, Chiba, 26 September 2010 through 30 September 2010
tmh_import_11_12_14. QC 201112202011-12-142011-12-142011-12-20Bibliographically approved