To improve speech recognition performance, feature transformation based on discriminant analysis has been widely used to reduce the redundant dimensions of acoustic features. Linear discriminant analysis (LDA) and heteroscedastic discriminant analysis (HDA) are often used for this purpose, and a generalization method for LDA and HDA, called power LDA (PLDA), has been proposed. However, these methods may result in an unexpected dimensionality reduction for multimodal data. It is important to preserve the local structure of the data when reducing the dimensionality of multimodal data. In this paper we introduce two methods, locality-preserving HDA and locality-preserving PLDA, to reduce dimensionality of multimodal data appropriately. We also propose an approximate calculation scheme to calculate sub-optimal projections rapidly. Experimental results show that the locality-preserving methods yield better performance than the traditional ones in speech recognition.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Makoto SAKAI, Norihide KITAOKA, Kazuya TAKEDA, "Acoustic Feature Transformation Based on Discriminant Analysis Preserving Local Structure for Speech Recognition" in IEICE TRANSACTIONS on Information,
vol. E93-D, no. 5, pp. 1244-1252, May 2010, doi: 10.1587/transinf.E93.D.1244.
Abstract: To improve speech recognition performance, feature transformation based on discriminant analysis has been widely used to reduce the redundant dimensions of acoustic features. Linear discriminant analysis (LDA) and heteroscedastic discriminant analysis (HDA) are often used for this purpose, and a generalization method for LDA and HDA, called power LDA (PLDA), has been proposed. However, these methods may result in an unexpected dimensionality reduction for multimodal data. It is important to preserve the local structure of the data when reducing the dimensionality of multimodal data. In this paper we introduce two methods, locality-preserving HDA and locality-preserving PLDA, to reduce dimensionality of multimodal data appropriately. We also propose an approximate calculation scheme to calculate sub-optimal projections rapidly. Experimental results show that the locality-preserving methods yield better performance than the traditional ones in speech recognition.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E93.D.1244/_p
Copy
@ARTICLE{e93-d_5_1244,
author={Makoto SAKAI, Norihide KITAOKA, Kazuya TAKEDA, },
journal={IEICE TRANSACTIONS on Information},
title={Acoustic Feature Transformation Based on Discriminant Analysis Preserving Local Structure for Speech Recognition},
year={2010},
volume={E93-D},
number={5},
pages={1244-1252},
abstract={To improve speech recognition performance, feature transformation based on discriminant analysis has been widely used to reduce the redundant dimensions of acoustic features. Linear discriminant analysis (LDA) and heteroscedastic discriminant analysis (HDA) are often used for this purpose, and a generalization method for LDA and HDA, called power LDA (PLDA), has been proposed. However, these methods may result in an unexpected dimensionality reduction for multimodal data. It is important to preserve the local structure of the data when reducing the dimensionality of multimodal data. In this paper we introduce two methods, locality-preserving HDA and locality-preserving PLDA, to reduce dimensionality of multimodal data appropriately. We also propose an approximate calculation scheme to calculate sub-optimal projections rapidly. Experimental results show that the locality-preserving methods yield better performance than the traditional ones in speech recognition.},
keywords={},
doi={10.1587/transinf.E93.D.1244},
ISSN={1745-1361},
month={May},}
Copy
TY - JOUR
TI - Acoustic Feature Transformation Based on Discriminant Analysis Preserving Local Structure for Speech Recognition
T2 - IEICE TRANSACTIONS on Information
SP - 1244
EP - 1252
AU - Makoto SAKAI
AU - Norihide KITAOKA
AU - Kazuya TAKEDA
PY - 2010
DO - 10.1587/transinf.E93.D.1244
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E93-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2010
AB - To improve speech recognition performance, feature transformation based on discriminant analysis has been widely used to reduce the redundant dimensions of acoustic features. Linear discriminant analysis (LDA) and heteroscedastic discriminant analysis (HDA) are often used for this purpose, and a generalization method for LDA and HDA, called power LDA (PLDA), has been proposed. However, these methods may result in an unexpected dimensionality reduction for multimodal data. It is important to preserve the local structure of the data when reducing the dimensionality of multimodal data. In this paper we introduce two methods, locality-preserving HDA and locality-preserving PLDA, to reduce dimensionality of multimodal data appropriately. We also propose an approximate calculation scheme to calculate sub-optimal projections rapidly. Experimental results show that the locality-preserving methods yield better performance than the traditional ones in speech recognition.
ER -