Speaker Verification Based on Wavelet Packets

Ganchev, Todor; Siafarikas, Mihalis; Fakotakis, Nikos

doi:10.1007/978-3-540-30120-2_38

Cited by 9 publications

(3 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, at frequency f 1 = 1000 Hz the DWPT resolution changes from 31.25 Hz to 62.5 Hz, and at frequency f 2 = 2250 Hz it changes from 62.5 Hz to 125 Hz. A set of 16 different WP trees, provisionally denoted as WP-0 to WP-15, was constructed each having a combination of different frequency resolutions in the frequency bands Ganchev et al (2004). In the present study, all potential solutions are explored in a systematic manner, and the best one among all WP trees is selected.…”

Section: Wavelet Packet Tree Designmentioning

confidence: 99%

See 1 more Smart Citation

Wavelet packet approximation of critical bands for speaker verification

Siafarikas

Ganchev

Fakotakis

et al. 2007

Int J Speech Technol

Self Cite

View full text Add to dashboard Cite

Exploiting the capabilities offered by the plethora of existing wavelets, together with the powerful set of orthonormal bases provided by wavelet packets, we construct a novel wavelet packet-based set of speech features that is optimized for the task of speaker verification. Our approach differs from previous wavelet-based work, primarily in the wavelet-packet tree design that follows the concept of critical bands, as well as in the particular wavelet basis function that has been used. In comparative experiments, we investigate several alternative speech parameterizations with respect to their usefulness for differentiating among human voices. The experimental results confirm that the proposed speech features outperform Mel-Frequency Cepstral Coefficients (MFCC) and previously used wavelet features on the task of speaker verification. A relative reduction of the equal error rate by 15%, 15% and 8% was observed for the proposed speech features, when compared to the wavelet packet features introduced by Farooq and Datta, the MFCC of Slaney, and the subband based cepstral coefficients of Sarikaya et al., respectively.

show abstract

Section: Wavelet Packet Tree Designmentioning

confidence: 99%

“…In two earlier works Ganchev et al 2004), the authors demonstrated two examples of successful wavelet packet-based speech features, fine-tuned for the task of speaker verification. In the present study, we fully develop the methodology for building such trees.…”

Section: Introductionmentioning

confidence: 98%

Wavelet packet approximation of critical bands for speaker verification

Siafarikas

Ganchev

Fakotakis

et al. 2007

Int J Speech Technol

Self Cite

View full text Add to dashboard Cite

show abstract

“…A better improvement can be achieved by using wavelet packet transform (WPT) [8][9][10][11]. WPT provides easy handling of the spectral content of speech signal, flexible utilization of the important frequency bands and a variable frequency resolution in each subband.…”

Section: Literature Reviewmentioning

confidence: 99%