Applying Discriminative Learning to Speaker Verification

Tales ImbiribaRafael MarinhoAdalbery CastroAldebaro Klautau

The Gaussian mixture model (GMM) is the main technique used in speaker recognition systems. However, in tasks other than speaker recognition, GMM is often outperformed by modern classifiers, such as support vector machines (SVM). This work seeks a better understanding of the reasons thatdiscriminative classifiers have not been as successful in speaker recognition, as in other applications. This is done by comparing GMM and a novel technique called discriminative GMM, which is similar to SVM in many aspects. Simulation results using the IME corpus show that DGMM can improve theperformance compared to GMM, and indicate that a proper model selection is essential to make SVM competitive in speaker verification.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: