Title :
Modified Linear Discriminant Analysis for Speech Recognition
Author :
Li, Xiao-Bing ; O´Shaughnessy, D.
Author_Institution :
INRS - Energy, Mater. & Telecommun., Montreal
Abstract :
In this paper, a new method for extracting discriminant features in automatic speech recognition (ASR), termed modified linear discriminant analysis (MLDA), is proposed. As a generalization of linear discriminant analysis (LDA), MLDA integrates the cluster information in each class by redefining the between-class scatter matrix based on the fact that many clusters exist in each state in hidden Markov model (HMM)-based ASR. Experimental results on TiDigits show that our presented MLDA clearly outperforms LDA and clustering-based linear discriminant analysis (CLDA), which was proposed for facial expression recognition, and about a 10% string error rate reduction (SERR) is found.
Keywords :
hidden Markov models; speech recognition; discriminant features; hidden Markov model; modified linear discriminant analysis; speech recognition; Automatic speech recognition; Covariance matrix; Data mining; Error analysis; Face recognition; Feature extraction; Hidden Markov models; Linear discriminant analysis; Scattering; Speech recognition;
Conference_Titel :
Electrical and Computer Engineering, 2007. CCECE 2007. Canadian Conference on
Conference_Location :
Vancouver, BC
Print_ISBN :
1-4244-1020-7
Electronic_ISBN :
0840-7789
DOI :
10.1109/CCECE.2007.400