Real-world speech/non-speech audio classification based on sparse representation features and GPCs

Ziqiang Shi; Jiqing Han; Tieran Zheng

doi:10.21437/interspeech.2011-625

Real-world speech/non-speech audio classification based on sparse representation features and GPCs

Ziqiang Shi^*
, Jiqing Han
, Tieran Zheng

^*Corresponding author for this work

School of Computer Science and Technology, Harbin Institute of Technology

Research output: Contribution to journal › Conference article › peer-review

1 Scopus citations

Abstract

A novel and robust approach for content based speech/nonspeech audio classification is proposed based on sparse representation (SR) features and Gaussian process classifiers (GPCs). The projections of the noise robust sparse representations for audio signals computed by L₁ -norm minimization are used as features. GPCs are used to learn and predict audio categories. Compare to the difficulties of Support Vector Machines (SVMs) in determining the hyperparameters, GPCs employ Bayesian selection criterion to estimate them. Experimental results on real-world audio datasets show that the SR features are more robust to audio variants than mel-frequency cepstral coefficients (MFCCs) and the proposed approach gives better performances than SVM.

Original language	English
Pages (from-to)	2401-2404
Number of pages	4
Journal	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
DOIs	https://doi.org/10.21437/interspeech.2011-625
State	Published - 2011
Externally published	Yes
Event	12th Annual Conference of the International Speech Communication Association, INTERSPEECH 2011 - Florence, Italy Duration: 27 Aug 2011 → 31 Aug 2011

Keywords

Audio classification
Gaussian process classifiers
Sparse representation
Speech discrimination

Access to Document

10.21437/interspeech.2011-625

Cite this

@article{1a3e4836c53f41e7b46a6fc96b054d53,

title = "Real-world speech/non-speech audio classification based on sparse representation features and GPCs",

abstract = "A novel and robust approach for content based speech/nonspeech audio classification is proposed based on sparse representation (SR) features and Gaussian process classifiers (GPCs). The projections of the noise robust sparse representations for audio signals computed by L1 -norm minimization are used as features. GPCs are used to learn and predict audio categories. Compare to the difficulties of Support Vector Machines (SVMs) in determining the hyperparameters, GPCs employ Bayesian selection criterion to estimate them. Experimental results on real-world audio datasets show that the SR features are more robust to audio variants than mel-frequency cepstral coefficients (MFCCs) and the proposed approach gives better performances than SVM.",

keywords = "Audio classification, Gaussian process classifiers, Sparse representation, Speech discrimination",

author = "Ziqiang Shi and Jiqing Han and Tieran Zheng",

year = "2011",

doi = "10.21437/interspeech.2011-625",

language = "英语",

pages = "2401--2404",

journal = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

issn = "2308-457X",

publisher = "International Speech Communication Association",

note = "12th Annual Conference of the International Speech Communication Association, INTERSPEECH 2011 ; Conference date: 27-08-2011 Through 31-08-2011",

}

TY - JOUR

T1 - Real-world speech/non-speech audio classification based on sparse representation features and GPCs

AU - Shi, Ziqiang

AU - Han, Jiqing

AU - Zheng, Tieran

PY - 2011

Y1 - 2011

N2 - A novel and robust approach for content based speech/nonspeech audio classification is proposed based on sparse representation (SR) features and Gaussian process classifiers (GPCs). The projections of the noise robust sparse representations for audio signals computed by L1 -norm minimization are used as features. GPCs are used to learn and predict audio categories. Compare to the difficulties of Support Vector Machines (SVMs) in determining the hyperparameters, GPCs employ Bayesian selection criterion to estimate them. Experimental results on real-world audio datasets show that the SR features are more robust to audio variants than mel-frequency cepstral coefficients (MFCCs) and the proposed approach gives better performances than SVM.

AB - A novel and robust approach for content based speech/nonspeech audio classification is proposed based on sparse representation (SR) features and Gaussian process classifiers (GPCs). The projections of the noise robust sparse representations for audio signals computed by L1 -norm minimization are used as features. GPCs are used to learn and predict audio categories. Compare to the difficulties of Support Vector Machines (SVMs) in determining the hyperparameters, GPCs employ Bayesian selection criterion to estimate them. Experimental results on real-world audio datasets show that the SR features are more robust to audio variants than mel-frequency cepstral coefficients (MFCCs) and the proposed approach gives better performances than SVM.

KW - Audio classification

KW - Gaussian process classifiers

KW - Sparse representation

KW - Speech discrimination

UR - https://www.scopus.com/pages/publications/84865726423

U2 - 10.21437/interspeech.2011-625

DO - 10.21437/interspeech.2011-625

M3 - 会议文章

AN - SCOPUS:84865726423

SN - 2308-457X

SP - 2401

EP - 2404

JO - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

JF - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

T2 - 12th Annual Conference of the International Speech Communication Association, INTERSPEECH 2011

Y2 - 27 August 2011 through 31 August 2011

ER -

Real-world speech/non-speech audio classification based on sparse representation features and GPCs

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this