论文信息 - A Novel Embedded Speaker Verification on System on Chip

A Novel Embedded Speaker Verification on System on Chip

This paper realizes a Text-Independent, Speaker Verification system on a System on Chip (SOC) platform. The system uses Mel-Frequency Cepstral Coefficients (MFCC) features with a Gaussian Mixture Model-Universal Background Model ¿GMM-UBM¿ speaker model. To deal with resource limitations, a new speaker-centric score normalization technique is introduced. This normalization technique results in a relative EER reduction of 44.9% compared to no normalization.

Jia Liu | Pengfei Mao

[1] Roland Auckenthaler,et al. Score Normalization for Text-Independent Speaker Verification Systems , 2000, Digit. Signal Process..

[2] Douglas A. Reynolds,et al. Speaker Verification Using Adapted Gaussian Mixture Models , 2000, Digit. Signal Process..

[3] Douglas A. Reynolds,et al. Speaker identification and verification using Gaussian mixture speaker models , 1995, Speech Commun..

[4] Chin-Hui Lee,et al. Maximum a posteriori estimation for multivariate Gaussian mixture observations of Markov chains , 1994, IEEE Trans. Speech Audio Process..

[5] Douglas A. Reynolds,et al. Robust text-independent speaker identification using Gaussian mixture speaker models , 1995, IEEE Trans. Speech Audio Process..

[6] Jia Liu,et al. A novel speech recognition system-on-chip , 2008, 2008 International Conference on Audio, Language and Image Processing.

[7] Douglas A. Reynolds,et al. Comparison of background normalization methods for text-independent speaker verification , 1997, EUROSPEECH.