4th International Conference on Spoken Language Processing

Philadelphia, PA, USA
October 3-6, 1996

Selective use of the Speech Spectrum and a VQGMM Method for Speaker Identification

Qiguang Lin, Ea-Ee Jan, ChiWei Che, Dong-Suk Yuk, James L. Flanagan

CAIP Center, Rutgers University, Busch Campus, NJ, USA

This paper describes two separate sets of speaker identification experiments. In the first set of experiments, the speech spectrum is selectively used for speaker identification. The results show that the higher portion of the speech spectrum contains more reliable idiosyncratic information on speakers than does the lower portion of equal bandwidth. In the second set of experiments, a vector-quantization based Gaussian mixture models (VQGMMs) is developed for text-independent speaker identification. The system has been evaluated in the recent speaker identification evaluation organized by NIST. In this paper, details of the system design are given and the evaluation results are presented.

Full Paper

Bibliographic reference.  Lin, Qiguang / Jan, Ea-Ee / Che, ChiWei / Yuk, Dong-Suk / Flanagan, James L. (1996): "Selective use of the speech spectrum and a VQGMM method for speaker identification", In ICSLP-1996, 2415-2418.