ISCA Archive Interspeech 2009
ISCA Archive Interspeech 2009

GMM kernel by Taylor series for speaker verification

Minqiang Xu, Xi Zhou, Beiqian Dai, Thomas S. Huang

Currently, approach of Gaussian Mixture Model combined with Support Vector Machine to text-independent speaker verification task has produced the stat-of-the-art performance. Many kernels have been reported for combining GMM and SVM.

In this paper, we propose a novel kernel to represent the GMM distribution by Taylor expansion theorem and itÂ’s regarded as the input of SVM. The utterance-specific GMM is represented as a combination of orders of Taylor series expansing at the means of the Gaussian components. Here we extract the distribution information around the means of the Gaussian components in the GMM as we can naturally assume that each mean position indicates a feature cluster in the feature space. And then the kernel computes the emsemble distance between orders of Taylor series.

Results of our new kernel on NIST speaker recognition evaluation (SRE) 2006 core task have been shown relative improvements of up to 7.1% and 11.7% in EER for male and female compared to K-L divergence based SVM system.

doi: 10.21437/Interspeech.2009-382

Cite as: Xu, M., Zhou, X., Dai, B., Huang, T.S. (2009) GMM kernel by Taylor series for speaker verification. Proc. Interspeech 2009, 1283-1286, doi: 10.21437/Interspeech.2009-382

  author={Minqiang Xu and Xi Zhou and Beiqian Dai and Thomas S. Huang},
  title={{GMM kernel by Taylor series for speaker verification}},
  booktitle={Proc. Interspeech 2009},