论文信息 - Real-Time Speaker Verification Based on GMM-UBM for PDA

Real-Time Speaker Verification Based on GMM-UBM for PDA

Personal digital assistant (PDA) is more and more popular due to its powerful functions. To protect the personal information of PDA, we adopt the technology of text-independent speaker verification (SV) for this kind of embedded system. In this paper, Gaussian mixture model (GMM) and universal background model (UBM) are addressed in detail. The optimization of mathematical computation of GMM-UBM guarantees the real-time performance of our SV system.

Qingyang Hong | Yan Chen | Caihong Zhang | XiaoYang Chen

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

[2] Sam Kwong,et al. A genetic classification method for speaker recognition , 2005, Eng. Appl. Artif. Intell..

[3] Alexander I. Rudnicky,et al. Pocketsphinx: A Free, Real-Time Continuous Speech Recognition System for Hand-Held Devices , 2006, 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings.

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

[5] Li Hui. Threshold Selection Based on TZ Normalization for Speaker Verification , 2005 .

[6] Yuet-Ming Lam,et al. FIXED-POINT IMPLEMENTATIONS OF SPEECH RECOGNITION SYSTEMS , 2002 .