On the use of PCA in GMM and AR-vector models for text independent speaker verification
2002 14th International Conference on Digital Signal Processing Proceedings. DSP 2002 (Cat. No.02TH8628) - Tập 2 - Trang 595-598 vol.2
Tóm tắt
This paper examines the role of the principal components analysis (PCA) on the performance of two classification systems for text independent speaker verification: the Gaussian mixture model (GMM) and the AR-vector model. The use of the PCA transform resulted in an improvement in the performance of the GMM for training times of 60 s and 30 s. However, the advantage of using PCA was not observed for the AR-vector model. For the case of 10 s training time, there was no benefit in using PCA even with GMM. In this situation, the AR-vector is superior for a 10 s test and worse for a 3 s test. In this latter case, however, all systems yielded error rates above 7%.
Từ khóa
#Principal component analysis #Hidden Markov models #Speech #Training data #Testing #Loudspeakers #Speaker recognition #Covariance matrix #Error analysis #TelephonyTài liệu tham khảo
10.1109/89.365379
reynolds, 1992, A Gaussian Mixture Modeling Approach to Text Independent Speaker Identification
10.1109/TASSP.1980.1163420
10.1006/dspr.1999.0361
10.1109/ICASSP.1992.226134
fukunaga, 1990, Introduction to Statistical Pattern Recognition
malayath, 2000, Data-driven Methods for Extracting Features From Speech
10.1016/0167-6393(95)00009-D
alcaim, 1992, Freqüência de ocorrência dos fonemas e listas de frases foneticamente balanceadas no Português falado no Rio de Janeiro, Revista da Sociedade Brasileira de Telecomunicaes, 7
10.1109/5.628714