Browse > Article

The Reduction or computation in MLLR Framework using PCA or ICA for Speaker Adaptation  

김지운 (인하대학교 전자공학과 DSP Lab)
정재호 (인하대학교 전자공학과 DSP Lab)
Abstract
We discuss how to reduce the number of inverse matrix and its dimensions requested in MLLR framework for speaker adaptation. To find a smaller set of variables with less redundancy, we adapt PCA (principal component analysis) and ICA (independent component analysis) that would give as good a representation as possible. The amount of additional computation when PCA or ICA is applied is as small as it can be disregarded. 10 components for ICA and 12 components for PCA represent similar performance with 36 components for ordinary MLLR framework. If dimension of SI model parameter is n, the amount of computation of inverse matrix in MLLR is proportioned to O(n⁴). So, compared with ordinary MLLR, the amount of total computation requested in speaker adaptation is reduced by about 1/81 in MLLR with PCA and 1/167 in MLLR with ICA.
Keywords
MLLR; Speaker adaptation; MLLR; PCA; ICA;
Citations & Related Records
연도 인용수 순위
  • Reference
1 A maximum-likelihood approach to stochastic matching for robust speech recognition /
[ A.Sankar;C.H.Lee ] / IEEE Trans. on Speech Audio Processing   DOI   ScienceOn
2 Probabilistic principal component analysis /
[ M.E.Tipping;C.M.Bishop ] / Journal of the Royal Statistical Society
3 Probabilistic PCS and ICA subspace mixture models for image segmentation /
[ D.Ridder;J.Kittler;R.P.W.Duin ] / The Eleventh British Machine Vision Conference
4 Joint maximum a posteriori adaptation of transformation and HMM parameters /
[ O.Siohan;C.Chesta;C.H.Lee ] / IEEE Trans. on Speech and Audio Processing   DOI   ScienceOn
5 /
[ A.Hyvarinen;J.Karhunen;E.Oja ] / Independent Component Analysis
6 A study on speaker adaptation of the parameters of continuous density hidden Markov models /
[ C.H.Lee;C.H.Lin;B.H.Juang ] / IEEE Trans. on Signal Processing   DOI   ScienceOn
7 Improved acoustic modelling for HMMs using linear transforms /
[ C.J.Leggetter ] / PhD Thesis, Univ. of Cambridge
8 Online hierarchical transformation of hidden Markov models for speech recognition /
[ J.T.Chien ] / IEEE Trans. on Speech and Audio Processing   DOI   ScienceOn
9 On-line adaptaion of hidden Markov models using incremental estimation algorithms /
[ V.Digalakis ] / Proc. 5th Eur. Conf. Speech Communication and Technology