Speaker Identification Using a Novel Combination of Sparse Representation and Gaussian Mixture Models

Abstract:

In recent years, sparse representation has become a very popular method for pattern recognition which could outperform the traditional methods. This paper presents a novel combination of sparse representation and traditional Gaussian mixture models. Each person’s dictionary or termed as subspace in this paper are learned using K-SVD algorithm while the entries are GMM mean matrixes union for each speaker. Then project the test utterance into each dictionary and finally make decision depending on the reconstruction errors. The experiments are conducted on the database collected in our anechoic chamber. The proposed approach results in different accuracy for different sparsity and dictionary size. In appropriate parameters, the accuracy can reach 98.5% which is fairly good.

• Periodical: Applied Mechanics and Materials (Volume 615)
• Pages: 265-269
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.615.265
• Online since: August 2014
• Authors: Yun Jie Ma
• Keywords: GMM, Learned Dictionary, Sparse Representation, Speaker Identification

References

[1] Kumar G S. Speaker recognition using GMM[J]. International Journal of Engineering Science and Technology, 2010, 2 (6): 2428-2436.

[2] Naseem Imran, Togneri Roberto, Bennamoun Mohammed. Sparse Representation for Speaker Identification[C]. International Conference on pattern Recognition, Aug 23-26, 2010, Istanbul: IEEE 2010: 4460-4463.

DOI: 10.1109/icpr.2010.1083

[3] Reynolds D A, Quatieri T F, Dunn R. Speaker verification using adapted Gaussian mixture models[J]. Dig. Signal Process, 2000, 10 (1-3): 19-41.

DOI: 10.1006/dspr.1999.0361

[4] Hairs B C, Sinha R. Speaker verification using Sparse Representation over KSVD Learned Dictionary[C]. National Conference on Communications , Feb 3-5, 2012, Kharagpur. 2012: 1-5.

DOI: 10.1109/ncc.2012.6176916

[5] Reynolds D A. Speaker identification and verification using Gaussian mixture speaker models[J]. Speech Communication, 1995, 17(1-2): 91-108.

DOI: 10.1016/0167-6393(95)00009-d

[6] Reynolds D A. Robust Text-independent speaker identification using Gaussian mixture speaker models[J]. IEEE trans. Speech and audio process, 1995, 13(1): 72-83.

DOI: 10.1109/89.365379

[7] Donoho D L. Compressed sensing[J]. IEEE trans. Inform. Theory, 2006, 52(4): 5406-5425.

[8] Candes E J, Wakin M B. An Introduction To Compressive Sampling. Signal Processing Magazine[J], IEEE, 2008, 25(2): 21-30.

DOI: 10.1109/msp.2007.914731

[9] Baraniuk R G. Compressive Sensing [J]. Signal Processing Magazine, IEEE, 2007, 24(4): 118-152.

[10] Aharon Michal, Elad Michael, Bruckstein Alfred. K-SVD: An Algorithm for Designing Overcomplete Dictionaries for Sparse Representation[J]. IEEE trans. Signal Processing, 2006, 54(11): 4311-4321.

DOI: 10.1109/tsp.2006.881199

[11] Rubinstein R, Peleg T, Elad M. Analysis K-SVD: A Dictionary-Learning Algorithm for the Analysis Sparse Model[J]. IEEE Trans. Signal Processing, 2013, 61(3): 661-667.

DOI: 10.1109/tsp.2012.2226445

[12] Barsi R, Jacobs D. Lambertian reflection and linear subspaces[J]. IEEE Trans. PAMI, 2003, 25(3): 218-233.