Spectral Subband Centroids as Complementary Features for Speaker Authentication

Thian, Norman Poh Hoon; Sanderson, Conrad; Bengio, Samy

doi:10.1007/978-3-540-25948-0_86

Cited by 22 publications

(15 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MFCCs were introduced in early 1980s for speech recognition and then adopted in speaker recognition. Even though various alternative features, such as spectral subband centroids (SSCs) [125,221] have been studied, the MFCCs seem to be difficult to beat in practice.…”

Section: Short-term Spectral Featuresmentioning

confidence: 99%

An overview of text-independent speaker recognition: From features to supervectors

Kinnunen¹,

Li²

2010

Speech Communication

1,221

706

View full text Add to dashboard Cite

This paper gives an overview of automatic speaker recognition technology, with an emphasis on text-independent recognition. Speaker recognition has been studied actively for several decades. We give an overview of both the classical and the state-of-the-art methods. We start with the fundamentals of automatic speaker recognition, concerning feature extraction and speaker modeling. We elaborate advanced computational techniques to address robustness and session variability. The recent progress from vectors towards supervectors opens up a new area of exploration and represents a technology trend. We also provide an overview of this recent development and discuss the evaluation methodology of speaker recognition systems. We conclude the paper with discussion on future directions.

show abstract

Section: Short-term Spectral Featuresmentioning

confidence: 99%

An overview of text-independent speaker recognition: From features to supervectors

Kinnunen¹,

Li²

2010

Speech Communication

1,221

706

View full text Add to dashboard Cite

show abstract

“…In [7] SSCs yielded comparable results to MFCCs in noise-free condition. Moreover, SSCs outperformed MFCCs under additive noise conditions with low signal-to-noise ratios.…”

Section: Discussionmentioning

confidence: 68%

“…-SSC(1) : linear frequency scale, non-overlapping rectangular filters -SSC(2) : mel frequency scale, non-overlapping rectangular filters -SSC(3) : mel frequency scale, overlapping triangular filters According to [7], mean subtraction helps SSC. We confirmed this experimentally and we apply it in all the three cases.…”

Section: Speaker Verification Resultsmentioning

confidence: 99%

“…One source for this disparity may arise from implemetation differences of the feature extraction, in particular, the channel compensation methods applied. In [7], mean subtraction was used with the SSC features, but no channel compensation was mentioned in conjunction with the MFCC features. Our MFCC front-end, on the other hand, includes RASTA filtering and utterance level mean/variance normalization to increase robustness.…”

Section: Discussionmentioning

confidence: 99%

“…SSCs are computed as the centroid frequencies of subband spectra and they give the locations of the local maxima of the power spectrum. SSCs have been used for speech recognition [4; 5], speaker recognition [7] and audio fingerprinting [6]. Recognition accuracy of SSCs is lower in noise-free conditions compared with MFCCs.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Speaker Verification with Adaptive Spectral Subband Centroids

Kinnunen

Zhang

Zhu

et al.

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Spectral subband centroids (SSC) have been used as an additional feature to cepstral coefficients in speech and speaker recognition. SSCs are computed as the centroid frequencies of subbands and they capture the dominant frequencies of the short-term spectrum. In the baseline SSC method, the subband filters are pre-specified. To allow better adaptation to formant movements and other dynamic phenomena, we propose to adapt the subband filter boundaries on a frame-by-frame basis using a globally optimal scalar quantization scheme. The method has only one control parameter, the number of subbands. Speaker verification results on the NIST 2001 task indicate that the selection of the parameter is not critical and that the method does not require additional feature normalization.

show abstract

A Score-Level Fusion Benchmark Database for Biometric Authentication

Poh

Bengio

2005

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Fusing the scores of several biometric systems is a very promising approach to improve the overall system's accuracy. Despite many works in the literature, it is surprising that there is no coordinated effort in making a benchmark database available. It should be noted that fusion in this context consists not only of multimodal fusion, but also intramodal fusion, i.e., fusing systems using the same biometric modality but different features, or same features but using different classifiers. Building baseline systems from scratch often prevents researchers from putting more efforts in understanding the fusion problem. This paper describes a database of scores taken from experiments carried out on the XM2VTS face and speaker verification database. It then proposes several fusion protocols and provides some state-of-the-art tools to evaluate the fusion performance.

show abstract

Spectral Subband Centroids as Complementary Features for Speaker Authentication

Cited by 22 publications

References 11 publications

An overview of text-independent speaker recognition: From features to supervectors

An overview of text-independent speaker recognition: From features to supervectors

Speaker Verification with Adaptive Spectral Subband Centroids

A Score-Level Fusion Benchmark Database for Biometric Authentication

Contact Info

Product

Resources

About