DocumentCode :
1386642
Title :
Optimized Discriminative Kernel for SVM Scoring and Its Application to Speaker Verification
Author :
Zhang, Shi-Xiong ; Mak, Man-Wai
Author_Institution :
Dept. of Electron. & Inf. Eng., Hong Kong Polytech. Univ., Hong Kong, China
Volume :
22
Issue :
2
fYear :
2011
Firstpage :
173
Lastpage :
185
Abstract :
The decision-making process of many binary classification systems is based on the likelihood ratio (LR) scores of test patterns. This paper shows that LR scores can be expressed in terms of the similarity between the supervectors (SVs) formed by stacking the mean vectors of Gaussian mixture models corresponding to the test patterns, the target model, and the background model. By interpreting the support vector machine (SVM) kernels as a specific similarity (or discriminant) function between SVs, this paper shows that LR scoring is a special case of SVM scoring and that most sequence kernels can be obtained by assuming a specific form for the similarity function of SVs. This paper further shows that this assumption can be relaxed to derive a new general kernel. The kernel function is general in that it is a linear combination of any kernels belonging to the reproducing kernel Hilbert space. The combination weights are obtained by optimizing the ability of a discriminant function to separate the positive and negative classes using either regression analysis or SVM training. The idea was applied to both high-and low-level speaker verification. In both cases, results show that the proposed kernels achieve better performance than several state-of-the-art sequence kernels. Further performance enhancement was also observed when the high-level scores were combined with acoustic scores.
Keywords :
Gaussian processes; Hilbert spaces; decision making; regression analysis; speaker recognition; support vector machines; Gaussian mixture models; SVM scoring; acoustic scores; binary classification systems; decision-making process; kernel Hilbert space; likelihood ratio scores; optimized discriminative kernel; regression analysis; similarity function; speaker verification; supervectors; support vector machine; Adaptation model; Computational modeling; Kernel; Measurement; Optimization; Speech; Support vector machines; Kernel optimization; sequence kernels; speaker verification; support vector machines; Algorithms; Artificial Intelligence; Classification; Discrimination Learning; Humans; Linear Models; Neural Networks (Computer); Normal Distribution; Pattern Recognition, Automated; Speech Recognition Software; Voice;
fLanguage :
English
Journal_Title :
Neural Networks, IEEE Transactions on
Publisher :
ieee
ISSN :
1045-9227
Type :
jour
DOI :
10.1109/TNN.2010.2090893
Filename :
5643156
Link To Document :
بازگشت