XiaoBing Li, Jinyu Li, and RenHua Wang
In this paper, Minimum Classification Error (MCE) method is extended to optimize both Linear Discriminant Analysis (LDA) transformation and the classification parameters for dimensionality reduction. Firstly, under the HMM-based Continuous Speech Recognition (CSR) framework, we use MCE criterion to optimize the conventional dimensionality reduction method, which uses LDA to transform standard MFCCs. Then, a new dimensionality reduction method is proposed. In the new method, the combination of Discrete Cosine Transform (DCT) and LDA, as used in the conventional method, is replaced by a single LDA transformation, which is optimized according to MCE criterion along with the classification parameters. Experimental results on TiDigits show that even when the feature dimension is reduced to 14, the performance of this new method is as good as that of the MCEtrained system using 39 dimension MFCCs. It also outperforms our MCE-optimized conventional dimensionality reduction method.
|Published in||Proc. ICASSP|