A Novel Bi-Group Particle Swarm Optimizer for Speaker Identification

Abstract:

The traditional training meshods of speaker codebook for speaker identification based on vector quantization are sensitive to the initial codebook parameters, and they often lead to a sub-optimal codebook in practice. To resolve this problem, this paper proposes a novel bi-group particle swarm optimizer (BPSO). It applies two sub-group particles with different particle update parameters simultaneously to explore the best speaker codebook, and the particles perform basic operations of particle swarm optimization (PSO) and conventional LBG algorithm in sequence, which can explore the solution space separately and search the local part in detail together. Information is exchanged when sub-groups are periodically shuffled and reorganized. Experimental results have demonstrated that the performance of BPSO is much better than that of LBG, fuzzy C-means (FCM), fast evolutionary programming (FEP), PSO, the impoved PSO algorithm consistently with higher correct identification rates and convergence rate. The dependence of the final codebook on the selection of the initial codebook is also reduced effectively.

• Periodical: Applied Mechanics and Materials (Volumes 380-384)
• Pages: 1298-1301
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.380-384.1298
• Online since: August 2013
• Authors: Li Ping Xue, Ying Long Yao, Hong Zhou, Ping Cai
• Keywords: Bi-Group, Particle Swarm Optimization (PSO), Shuffled Frog Leaping Algorithm, Speaker Identification, Vector Quantization

References

[1] Yu K, Mason J, Oglesby J. Speaker recognition using hidden Markov models, dynamic time warping and vector quantisation[J]. IEE Proc., Vis. Image Signal Processing, 1995, 142(5): 313–318.

DOI: 10.1049/ip-vis:19952144

[2] Soong F K, Rosenberg A E, Rabiner L. R, et al. A vector quantization approach to speaker recognition[A]. In Proc. IEEE ICASSP'85 [C]. Florida: IEEE Press, 1985. 387–390.

DOI: 10.1109/icassp.1985.1168412

[3] Tran D, Wagner M, Van L T. A proposed decision rule for speaker recognition based on fuzzy C-Means clustering [A]. In 5th Int Conf on Spoken Language Processing[C]. Sydney: ASSTA Press, 1998. 755-758.

DOI: 10.21437/icslp.1998-407

[4] Wang Jin-ming, Li En-bo. Research and implementation of speaker identification based on GA/VQ[J]. Journal of PLA University of Science and Technology (Natural Science Edition), 2005, 6(3): 214–218.

[5] Hu Heng-tao, Long Jian-hong. Speaker identification using fussy C-means clustering algorithm based on improved ant agony algorithm. Journal of Sichuan University (Natural Science Edition), 2007, 44(3): 543-547. (in Chinese).

[6] Kennedy J, Eberhart R. Particle swarm optimization [A]. In Proc. IEEE Int Conf on Neural Networks 1995[C]. Piscataway: IEEE Press, 1995. 1942-(1948).

[7] Eusuffm M, Lansey K E. Optimization of Water Distribution Network Design Using Shuffled Particle Leaping Algorithm[J]. Journal of Water Resources Planning and Management, 2003, 129 (3): 210-225.

DOI: 10.1061/(asce)0733-9496(2003)129:3(210)

[8] Xue Li-ping, Yin Jun-xun, Ji Zhen. Speaker recognition based on particle swarm optimization and fuzzy clustering Analysis [J]. Journal of Shenzhen University Science and Engineering. 2008, 25(3): 178-183.

[9] Xue Li-ping, Yin Jun-xun, Zhou Jiarui, et al. Speaker Identification Based on Particle-Pair Cooperative Optimizer [J]. Acta Electronics Sinica, 2009, 37(1): 207-211.

[10] Yao X, Liu Y, Lin G M. Evolutionary Programming Made Faster [J]. IEEE Transactions on Evolutionary Computation, 1999, 3 (2): 82-102.

DOI: 10.1109/4235.771163

[11] Garofolo J S, Lamel L F. TIMIT Acoustic-Phonetic Continuous Speech Corpus [DB/CD]. http: /www. ldc. upenn. edu/Catalog/, 2012-10.

DOI: 10.6028/nist.ir.4930