Title :
Data-driven, nonlinear, formant-to-acoustic mapping for ASR
Author :
Jackson, P.J.B. ; Lo, B.-H. ; Russell, M.J.
Author_Institution :
Dept. of Electron., Electr. & Comput. Eng., Univ. of Birmingham, UK
fDate :
6/20/2002 12:00:00 AM
Abstract :
With a view to using an articulatory representation in automatic recognition of conversational speech, two nonlinear methods for mapping from formants to short-term spectra were investigated: multilayered perceptrons (MLPs), and radial basis function (RBF) networks. Five schemes for dividing the TIMIT data according to their phone class were tested. The r.m.s. error of the RBF networks was 10%, less than that of the MLP, and the scheme based on discrete articulatory regions gave the greatest improvements over a single network
Keywords :
multilayer perceptrons; radial basis function networks; speech recognition; ASR; MLP; RBF networks; RMS error; TIMIT data; articulatory representation; automatic speech recognition; conversational speech; formant-to-acoustic mapping; multilayered perceptrons; nonlinear methods; phone class; radial basis function networks; short-term spectra;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20020436
