8th Annual Conference of the International Speech Communication Association

Antwerp, Belgium
August 27-31, 2007

Cluster Adaptive Training Weights as Features in SVM-Based Speaker Verification

Hao Yang (1), Yuan Dong (2), Xianyu Zhao (2), Jian Zhao (1), Liang Lu (1), Haila Wang (2)

(1) Beijing University of Posts & Telecommunications, China
(2) France Telecom R&D Beijing, China

In this paper, we propose the use of cluster adaptive training (CAT) weights as features in support vector machine (SVM) based text-independent verification task. The speaker utterance is characterized by a vector of cluster weights, which are extracted during the cluster adaptive training process. The effects of the number of classes, which are obtained by partitioning the components of the model, and the number of clusters on the verification performance are investigated. To remove session variability due to influences of microphone, environment, etc, Nuisance Attribute Projection (NAP) is also evaluated. Experimental results in a NIST SRE 2006 task show that this CAT weights SVM system achieves comparable performance to a state-of-the-art cepstral GMM-UBM verification system, and their fusion can give further performance gains.

Full Paper

Bibliographic reference.  Yang, Hao / Dong, Yuan / Zhao, Xianyu / Zhao, Jian / Lu, Liang / Wang, Haila (2007): "Cluster adaptive training weights as features in SVM-based speaker verification", In INTERSPEECH-2007, 2013-2016.