Share Email Print
cover

Proceedings Paper

Design of speaker recognition system based on artificial neural network
Author(s): Yanhong Chen; Li Wang; Han Lin; Jinlong Li
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Speaker recognition is to recognize speaker’s identity from its voice which contains physiological and behavioral characteristics unique to each individual. In this paper, the artificial neural network model, which has very good capacity of non-linear division in characteristic space, is used for pattern matching. The speaker's sample characteristic domain is built for his mixed voice characteristic signals based on Kmeanlbg algorithm. Then the dimension of the inputting eigenvector is reduced, and the redundant information is got rid of. On this basis, BP neural network is used to divide capacity area for characteristic space nonlinearly, and the BP neural network acts as a classifier for the speaker. Finally, a speaker recognition system based on the neural network is realized and the experiment results validate the recognition performance and robustness of the system.

Paper Details

Date Published: 15 October 2012
PDF: 7 pages
Proc. SPIE 8420, 6th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optical System Technologies for Manufacturing and Testing, 84200U (15 October 2012); doi: 10.1117/12.970642
Show Author Affiliations
Yanhong Chen, Southwest Jiaotong Univ. (China)
Li Wang, Southwest Jiaotong Univ. (China)
Han Lin, Southwest Jiaotong Univ. (China)
Jinlong Li, Southwest Jiaotong Univ. (China)


Published in SPIE Proceedings Vol. 8420:
6th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optical System Technologies for Manufacturing and Testing
Xiangdi Lin; Yoshiharu Namba; Tingwen Xing, Editor(s)

© SPIE. Terms of Use
Back to Top