Suzan A. Mahmood1,   Loay E. George2

1 College of Science, University of Sulaimani, 2 College of Science, University of Baghdad

DOI: https://doi.org/10.17656/jzs.10181

Abstract

In this paper, a neural-based system for speaker identification is investigated. The linear predictive

coefficients have been used as a set of descriptors for the speaker data. This set of descriptors is fed to a back

propagation neural network for the purpose of speaker identification. Also, it is found that the process of

segmenting the audio data into slices has great effect on the identification results, the number and length of

slice has major influence on the success of the identification process.

Keywords: Leaner prediction coding, Speaker identification, backpropagation NN.

References

1.  Marek Szala “Two-level pattern recognition in a class of knowledge-based systems” , Knowledge-Based Systems, 2002, 15(1), 95-101.

2.  Amarin Deemagarn, Asanee Kawtrakul  “ Thai Connected Digit Speech Recognition     Using Hidden Markov Models”, International Conference Speech and Computer, 2004,     9(1),209-213. 

3.  Choubassi, M. M. El,  El Khoury, H. E., Jabra Alagha, C. E.,Skaf, J.A.and Al-Alaoui, M.A. “Arabic Speech Recognition Using Recurrent Neural Networks”, IEEE Intl.    Symp. Signal Processing and Information Technology ISSPIT, 2003, 3(1), 336-340.

4.  Campbell,J. P.Jr. , ”Speaker Recognition”,PROCEEDINGS OF THE IEEE,1997, 85(9),     284-366.

5.  Farell, K., Mammone,R & Assaleh, K.“Speaker Recognition Using Neural Networks and Conventional Classifiers” IEEE Transactions on Speech and Audio Processing, 1994,     2(1), Part II, 194-205.

6.  Chularat Tanprasert, ” Text-dependent Speaker Identification Using Neural Network On  Distinctive Thai Tone Marks”, NECTEC Technical Journal, 2000,1(6), 123-129. 

View Full Article