Speaker Adaptation Based on PARAFAC2 of Transformation Matrices for Continuous Speech Recognition

Yongwon JEONG; Sangjun LIM; Young Kuk KIM; Hyung Soon KIM

doi:10.1587/transinf.E96.D.2152

IEICE Transactions on Information and Systems

Online ISSN : 1745-1361
Print ISSN : 0916-8532

Regular Section

Speaker Adaptation Based on PARAFAC2 of Transformation Matrices for Continuous Speech Recognition

Yongwon JEONG, Sangjun LIM, Young Kuk KIM, Hyung Soon KIM

Author information

Keywords: maximum likelihood linear regression, parallel factor analysis, PARAFAC2, speaker adaptation, speech recognition

JOURNAL FREE ACCESS

2013 Volume E96.D Issue 9 Pages 2152-2155

DOI https://doi.org/10.1587/transinf.E96.D.2152

Details

Abstract

We present an acoustic model adaptation method where the transformation matrix for a new speaker is given by the product of bases and a weight matrix. The bases are built from the parallel factor analysis 2 (PARAFAC2) of training speakers' transformation matrices. We perform continuous speech recognition experiments using the WSJ0 corpus.

Content from these authors

Favorites & Alerts

Corresponding author

Register with J-STAGE for free!