This paper presents the performance of the AR-Vector with Cepstral Mean Subtraction (CMS) used to... more This paper presents the performance of the AR-Vector with Cepstral Mean Subtraction (CMS) used to compensate the distortions caused by distinct telephone channels. The performance obtained with the use of CMS is compared with a system without compensation. With ¼× of speech signal used for training and ¿¼× used for testing, the error rate without channel normalization is around ¾ ¾% against the ½ % achieved with CMS. For ½¼× testing time, the error rate dropped from ¼% to ¿ ¼% when using CMS. For the lowest testing time (¿×), the error rate of the AR-Vector is close to ½ % regardless the use or not of the normalization technique. Although there is a clear improvement in performance when using CMS, it is not of major significance. This leads to the conclusion that the AR-Vector classification system is somewhat robust to channel distortion, especially as the testing time decreases.
This paper examines the role of the principal components analysis (PCA) on the performance of two... more This paper examines the role of the principal components analysis (PCA) on the performance of two classification systems for text independent speaker verification: the Gaussian mixture model (GMM) and the AR-vector model. The use of the PCA transform resulted in an improvement in the performance of the GMM for training times of 60 s and 30 s. However, the advantage of using PCA was not observed for the AR-vector model. For the case of 10 s training time, there was no benefit in using PCA even with GMM. In this situation, the AR-vector is superior for a 10 s test and worse for a 3 s test. In this latter case, however, all systems yielded error rates above 7%.
Guilherme Stolle Paixão e Casarões * O presente artigo constitui um dos resultados da pesquisa "I... more Guilherme Stolle Paixão e Casarões * O presente artigo constitui um dos resultados da pesquisa "Impeachment: crise e transição política", que conta com o apoio do CNPq.
This paper presents the performance of the AR-Vector with Cepstral Mean Subtraction (CMS) used to... more This paper presents the performance of the AR-Vector with Cepstral Mean Subtraction (CMS) used to compensate the distortions caused by distinct telephone channels. The performance obtained with the use of CMS is compared with a system without compensation. With ¼× of speech signal used for training and ¿¼× used for testing, the error rate without channel normalization is around ¾ ¾% against the ½ % achieved with CMS. For ½¼× testing time, the error rate dropped from ¼% to ¿ ¼% when using CMS. For the lowest testing time (¿×), the error rate of the AR-Vector is close to ½ % regardless the use or not of the normalization technique. Although there is a clear improvement in performance when using CMS, it is not of major significance. This leads to the conclusion that the AR-Vector classification system is somewhat robust to channel distortion, especially as the testing time decreases.
This paper examines the role of the principal components analysis (PCA) on the performance of two... more This paper examines the role of the principal components analysis (PCA) on the performance of two classification systems for text independent speaker verification: the Gaussian mixture model (GMM) and the AR-vector model. The use of the PCA transform resulted in an improvement in the performance of the GMM for training times of 60 s and 30 s. However, the advantage of using PCA was not observed for the AR-vector model. For the case of 10 s training time, there was no benefit in using PCA even with GMM. In this situation, the AR-vector is superior for a 10 s test and worse for a 3 s test. In this latter case, however, all systems yielded error rates above 7%.
Guilherme Stolle Paixão e Casarões * O presente artigo constitui um dos resultados da pesquisa "I... more Guilherme Stolle Paixão e Casarões * O presente artigo constitui um dos resultados da pesquisa "Impeachment: crise e transição política", que conta com o apoio do CNPq.
Uploads
Papers by Charles Lima