Title | ||
---|---|---|
A novel alternative hypothesis characterization using kernel classifiers for LLR-Based speaker verification |
Abstract | ||
---|---|---|
In a log-likelihood ratio (LLR)-based speaker verification system, the alternative hypothesis is usually ill-defined and hard to characterize a priori, since it should cover the space of all possible impostors. In this paper, we propose a new LLR measure in an attempt to characterize the alternative hypothesis in a more effective and robust way than conventional methods. This LLR measure can be further formulated as a non-linear discriminant classifier and solved by kernel-based techniques, such as the Kernel Fisher Discriminant (KFD) and Support Vector Machine (SVM). The results of experiments on two speaker verification tasks show that the proposed methods outperform classical LLR-based approaches. |
Year | DOI | Venue |
---|---|---|
2006 | 10.1007/11939993_53 | ISCSLP |
Keywords | Field | DocType |
new llr measure,conventional method,novel alternative hypothesis characterization,classical llr-based approach,kernel fisher discriminant,llr-based speaker verification,llr measure,support vector machine,speaker verification task,alternative hypothesis,speaker verification system,kernel-based technique,kernel classifier,log likelihood ratio | Kernel (linear algebra),Alternative hypothesis,Likelihood-ratio test,Pattern recognition,Computer science,Support vector machine,A priori and a posteriori,Speech recognition,Artificial intelligence,Linear discriminant analysis,Kernel method,Classifier (linguistics) | Conference |
Volume | ISSN | ISBN |
4274 | 0302-9743 | 3-540-49665-3 |
Citations | PageRank | References |
0 | 0.34 | 8 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yi-Hsiang Chao | 1 | 40 | 6.39 |
Hsin-min Wang | 2 | 1201 | 129.62 |
Ruei-Chuan Chang | 3 | 267 | 56.19 |