SimNet Neural Network: An Application to Speaker Identification

Author

David Lee Enke, Missouri University of Science and TechnologyFollow
Hyoung-Koo Lee, Missouri University of Science and TechnologyFollow
A. M. Ozbayoglu
A. Thammano
Cihan H. Dagli, Missouri University of Science and TechnologyFollow

Abstract

Although the initial processes involved in speaker identification has gained considerable interest in recent years, the role of classification and recognition of the processed signal has generated less attention. In fact, when neural networks are used for classification, a standard backpropagation architecture is usually implemented. The following paper implements power spectrum values as inputs and a novel neural network, SimNet, for the classification and recognition stage of a speaker identification system. The performance of the SimNet on a database of 20 speakers, uttering the single word `Hi,' has produced a classification and recognition accuracy of 91.3%.

Recommended Citation

D. L. Enke et al., "SimNet Neural Network: An Application to Speaker Identification," Intelligent Engineering Systems Through Artificial Neural Networks, vol. 5, pp. 69 - 76, Dec 1995.

Department(s)

Electrical and Computer Engineering

Second Department

Nuclear Engineering and Radiation Science

Third Department

Engineering Management and Systems Engineering

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2024 The Authors, All rights reserved.

Publication Date

01 Dec 1995