Bispectrum features for robust speaker identification

Wenndt, Stanley J.; Shamsunder, S.

doi:10.1109/icassp.1997.596132

Cited by 7 publications

(6 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For each reference, bispectral features (using direct FFT method [6]) were computed by summing up values in the 0<f1<0.1 slice along the f2 axis. Energy, ZCR, autocorrelation, and high-tolow frequency ratio were also computed for each reference.…”

Section: Resultsmentioning

confidence: 99%

“…Moreover, third order frequency domain polyspectral measure is sensitive only to non-Gaussian entities. Hence, non-Gaussian speech mixed with Gaussian noise is expected to be better represented through bispectral analysis [6].…”

Section: Bispectral Feature Extractionmentioning

confidence: 99%

“…Nonlinear nature of speech has been analyzed in [4]- [5], for example, and, in particular, the possibility of using bispectrum for speaker identification was studied in [6] with excellent results.…”

Section: Introductionmentioning

confidence: 98%

See 2 more Smart Citations

Word boundary detection through frame classification using bispectral analysis

Umakanthan

Gopalan

2012

2012 IEEE International Conference on Electro/Information Technology

View full text Add to dashboard Cite

This paper presents a word boundary detection technique based on frame classification using the nonlinear characteristics of speech. Bispectral analysis was used to classify speech frames into voiced, unvoiced and noise segments. To improve classification accuracy, bispectral features were combined with other features such as short time energy, zero-crossing rate, autocorrelation and high-to-low frequency ratio. Experimental results indicate that classification error decreases when bispectrum is combined with other features. Thus bispectral features can be used as supplementary to augment simple time domain features for demarcating word boundaries in speech. Validation of results was carried out by manual verification.Index Terms-bispectrum; word boundary detection;

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Bispectral Feature Extractionmentioning

confidence: 99%

See 1 more Smart Citation

Word boundary detection through frame classification using bispectral analysis

Umakanthan

Gopalan

2012

2012 IEEE International Conference on Electro/Information Technology

View full text Add to dashboard Cite

show abstract

“…In resent years, higher-order statistics (HOS) have received increasing interest, especially in the form of higher-order cumulants and spectra. It has been used in particular to extract information, and to detect and characterize non-linear properties in signals [12][13] [14]. Furthermore, it has been found that the third-order spectrum, called bispectrum, has the ability to hold the information both in phase and in amplitude.…”

Section: A Review Of Bispectrummentioning

confidence: 99%

Feature extraction using surrounding-line integral bispectrum for radar emitter signal

Chen

Jin

2008

2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence)

View full text Add to dashboard Cite

In everchanging threat emitter environment, specific emitter identification (SEI) technology extracts subtle but persistent features from received pulse signal to create a fingerprint unique to a specific radar. Unlike conventional five parameters deinterleaving algorithm, which can be grossly ambiguous for radar emitter sorting, the SEI technology provides hardware specific identification. In this paper, we propose an approach for extracting unintentional phase modulation features caused by oscillator based on surrounding-line integral bispectrum. The quantitative features, i.e. bispectra entropy, waveform entropy and mean of surrounding-line integrated bispectra, is extracted using entropy-like function to reveal the subtle difference between emitters. Computer simulations show that how the phase-noise-induced signal changes analysis based on bispectrum approach can be used to determine which of emitters transmitted a pulse signal.

show abstract

“…Cependant, le succès de l'utilisation des coefficients cepstraux a été freiné par des La méthode du bispectre (statistiques d'ordre supérieur) a été introduite par Wenndt et al [116]. Dans la situation où les données d'apprentissage sont propres et les données testées sont bruitées, les paramètres de bispectre donnent de meilleurs résultats que ceux du cepstre.…”

Section: Paramètres Spectro-temporels Et Prosodiquesunclassified

Discrimination parole/musique et étude de nouveaux paramètres et modèles pour un système d'identification du locuteur dans le contexte de conférences téléphoniques /

Ezzaidi¹

2002

View full text Add to dashboard Cite

Bispectrum features for robust speaker identification

Cited by 7 publications

References 5 publications

Word boundary detection through frame classification using bispectral analysis

Word boundary detection through frame classification using bispectral analysis

Feature extraction using surrounding-line integral bispectrum for radar emitter signal

Discrimination parole/musique et étude de nouveaux paramètres et modèles pour un système d'identification du locuteur dans le contexte de conférences téléphoniques /

Contact Info

Product

Resources

About