Research on MLLR based speaker recognition algorithm

被引：0
作者：
Tsinghua National Laboratory for Information Science and Technology , Department of Electronic Engineering, Tsinghua University, Beijing 100084, China [1 ]
机构：
[1] Tsinghua National Laboratory for Information Science and Technology (TNList), Department of Electronic Engineering, Tsinghua University
来源：
Zidonghua Xuebao Acta Auto. Sin. | 2009年 / 5卷 / 546-550期
关键词：
Channel compensation; Maximum likelihood linear regression (MLLR); Speaker recognition; Support vector machine (SVM);
D O I：
10.3724/SP.J.1004.2009.00546
中图分类号：
学科分类号：
摘要：
This paper uses the maximum likelihood linear regression (MLLR) as feature for text-independent speaker recognition algorithm. We introduce a universal background model (UBM) based MLLRSV-SVM algorithm first, and then extend the algorithm to multi-class for improvement. After channel compensation, in terms of the NIST 2006 SRE lconv4w-lconv4w/mic corpus, the MLLR based system is comparable with and complementary of the state of the art systems. The performance is greatly improved by simply linear fusion.
引用
收藏
页码：546 / 550
页数：4
相关论文
共 13 条
  • [1] Reynolds D.A., Quatieri T.F., Dunn R.B., Speaker verification using adapted Gaussian mixture models, Digital Signal Processing, 10, 1-3, pp. 19-41, (2000)
  • [2] Campbell W.M., Sturim D.E., Reynolds D.A., Support vector machines using GMM supervectors for speaker verification, IEEE Signal Processing Letters, 13, 5, pp. 308-311, (2006)
  • [3] Castaldo F., Colibro D., Dalmasso E., Laface P., Vair C., Compensation of nuisance factors for speaker and language recognition, IEEE Transactions on Audio, Speech, and Language Processing, 15, 7, pp. 1969-1978, (2007)
  • [4] Vair C., Colibro D., Castaldo F., Daimasso F., Laface P., Channel factors compensation in model and feature domain for speaker recognition, Proceedings of IEEE Odyssey: The Speaker and Language Recognition Workshop, pp. 1-6, (2006)
  • [5] Campbell W.M., Sturim D.E., Reynolds D.A., Solomonoff A., SVM based speaker verification using a GMM supervector kernel and NAP variability compensation, Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing, pp. 97-100, (2006)
  • [6] Guo W., Dai L.-R., Wang R.-H., Speaker verification based on improved updates to the SVM, Journal of Tsinghua University (Science and Technology), 48, z1, pp. 704-707, (2008)
  • [7] Stolcke A., Ferrer L., Kajarekar S., Shriberg E., Venkataraman A., MLLR transforms as features in speaker recognition, Proceedings of the 9th European Conference on Speech Communication and Technology, pp. 2425-2428, (2005)
  • [8] Karam Z.N., Campbell W.M., A new kernel for SVM MLLR based speaker recognition, Proceedings of the 8th Conference in the Annual Series of Interspeech Events and the 10th Biennial Eurospeech Conference, pp. 290-293, (2007)
  • [9] Pavel M., Petr S., Jan C., Pavel C., Phono-tactic language identification using high quality phoneme recognition, Proceedings of the 9th European Conference on Speech Communication and Technology, pp. 2237-2240, (2005)
  • [10] Bian Z.-Q., Zhang X.-G., Pattern Recognition, (1999)
12