Blind Stochastic Feature Transformation for Channel Robust Speaker Verification

Yiu, Kwok-Kwong; Mak, Man‐Wai; Cheung, M. C.; Kung, Sun‐Yuan

doi:10.1007/s11265-006-4174-4

Cited by 7 publications

(10 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…is the phonetic class of frame t. For the acoustic GMM system, we applied feature transformation [26] and short-time Gaussianization [27] to reduce the effect of channel distortion. Then, acoustic scores S GMM were computed based on GMM-UBM framework [1]:…”

Section: Scoring Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Speaker Verification via High-Level Feature Based Phonetic-Class Pronunciation Modeling

Zhang

Mak

Meng

2007

IEEE Trans. Comput.

View full text Add to dashboard Cite

Abstract-It has been shown recently that the pronunciation characteristics of speakers can be represented by articulatory feature-based conditional pronunciation models (AFCPMs). However, the pronunciation models are phoneme-dependent, which may lead to speaker models with low discriminative power when the amount of enrollment data is limited. This paper proposes to mitigate this problem by grouping similar phonemes into phonetic classes and representing background and speaker models as phonetic-class dependent density functions. Phonemes are grouped by (1) vector quantizing the discrete densities in the phoneme-dependent universal background models, (2) using the phone properties specified in the classical phoneme tree, or (3) combining vector quantization and phone properties. Evaluations based on 2000 NIST SRE show that this phonetic-class approach effectively alleviates the data spareness problem encountered in conventional AFCPM, which results in better performance when fused with acoustic features.

show abstract

Section: Scoring Methodsmentioning

confidence: 99%

“…The MFCCs and delta MFCCs were concatenated to form 38-dimensional feature vectors. Cepstral mean subtraction (CMS), fast blind stochastic features transformation (fBSFT) [26], [3] and short-time Gaussianization (STG) [27] were applied to the MFCCs to remove channel effects.…”

Section: A Speech Corpora and Speech Featuresmentioning

confidence: 99%

Speaker Verification via High-Level Feature Based Phonetic-Class Pronunciation Modeling

Zhang

Mak

Meng

2007

IEEE Trans. Comput.

View full text Add to dashboard Cite

show abstract

“…It has been shown that both mean normalization and bandpass filtering can minimize the filtering effect of linear channels [7,8]. However, these techniques may cause performance degradation when both training and recognition are derived from the same acoustic environment [2]. The second type of blind compensation transforms the distorted features such that acoustic environments have minimum effect on the distribution of the transformed features.…”

Section: Blind Compensationmentioning

confidence: 99%

“…The goal of channel compensation is to achieve performance approaching that of a "matched condition" system. Channel compensation can be applied in feature space [1,2], model space [3,4] or score space [5]. One advantage of featurespace compensation is that it is not necessary to modify the speaker models after training.…”

Section: Introductionmentioning

confidence: 99%

Probabilistic Feature Transformation for Channel Robust Speaker Verification

Mak

Yiu

2006

2006 16th IEEE Signal Processing Society Workshop on Machine Learning for Signal Processing

View full text Add to dashboard Cite

Feature transformation plays an important role in robust speaker verification over telephone networks. This paper compares several feature transformation techniques and evaluates their verification performance and computation time under the 2000 NIST speaker recognition evaluation protocol. Techniques compared include feature mapping (FM), stochastic feature transformation (SFT), and blind stochastic feature transformation (BSFT). The paper proposes a probabilistic feature mapping (PFM) in which the mapped features depend not only on the top-1 decoded Gaussian but also on the posterior probabilities of other Gaussians in the root model. The paper also proposes speeding up the computation of PFM and BSFT parameters by considering the top few Gaussians only. Results show that PFM performs slightly better than FM and that the fast approach can reduce computation time substantially. Among the approaches investigated, the fast BSFT is found to have the highest potential for robust speaker verification over telephone networks because it can achieve good performance without any a priori knowledge of the communication channel. It was also found that fusion of the scores derived from systems using BSFT and PFM can reduce the error rate further.

show abstract

“…Channel compensation can be applied in feature space [1,2], model space [3,4] or score space [5]. One advantage of feature-space compensation is that it is not necessary to modify the speaker models after training.…”

Section: Introductionmentioning

confidence: 99%

Probabilistic feature-based transformation for speaker verification over telephone networks

2007

View full text Add to dashboard Cite

Feature transformation aims to reduce the effects of channel-and handset-distortion in telephone-based speaker verification. This paper compares several feature transformation techniques and evaluates their verification performance and computation time under the 2000 NIST speaker recognition evaluation protocol. Techniques compared include feature mapping (FM), stochastic feature transformation (SFT), blind stochastic feature transformation (BSFT), feature warping (FW), and short-time Gaussianization (STG). The paper proposes a probabilistic feature mapping (PFM) in which the mapped features depend not only on the top-1 decoded Gaussian but also on the posterior probabilities of other Gaussians in the root model. The paper also proposes speeding up the computation of PFM and BSFT parameters by considering the top few Gaussians only. Results show that PFM performs slightly better than FM and that the fast approach can reduce computation time substantially. Among the approaches investigated, the fast BSFT (fBSFT) strikes a good balance between computational complexity and error rates, and FW and STG are the best in terms of error rates but with higher computational complexity. It was also found that fusion of the scores derived from systems using fBSFT and STG can reduce the error rate further. This study advocates that fBSFT, FW, and STG have the highest potential for robust speaker verification over telephone networks because they achieve good performance without any a priori knowledge of the communication channel.

show abstract

Blind Stochastic Feature Transformation for Channel Robust Speaker Verification

Cited by 7 publications

References 14 publications

Speaker Verification via High-Level Feature Based Phonetic-Class Pronunciation Modeling

Speaker Verification via High-Level Feature Based Phonetic-Class Pronunciation Modeling

Probabilistic Feature Transformation for Channel Robust Speaker Verification

Probabilistic feature-based transformation for speaker verification over telephone networks

Contact Info

Product

Resources

About