Speaker verification using minimum verification error training

Rosenberg, A. E.; Siohan, Olivier; Parathasarathy, Srinivasan

doi:10.1109/icassp.1998.674378

Cited by 30 publications

(21 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The minimum verification error (MVE) training method [6] is modified to fit our requirement that only mis-verified training samples should be considered. We call it the minimum verification squared-error (MVSE) adaptation strategy.…”

Section: Minimum Verification Squared-error (Mvse) Adaptation Strategymentioning

confidence: 99%

“…Although several discriminative training methods [4], such as the minimum classification error (MCE) training method [5], the minimum verification error (MVE) training method [6], and the maximum mutual information (MMI) training method [7], have been proposed, they tend to over-train a model if the amount of training data is insufficient. In contrast, the DFA framework is based on the minimum verification squared-error (MVSE) adaptation strategy, which is modified from the MVE training method.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Discriminative Feedback Adaptation for GMM-UBM Speaker Verification

Chao

Tsai

Wang

2008

2008 6th International Symposium on Chinese Spoken Language Processing

View full text Add to dashboard Cite

The GMM-UBM system is the current state-of-the-art approach for text-independent speaker verification. The advantage of the approach is that both target speaker model and impostor model (UBM) have generalization ability to handle "unseen" acoustic patterns. However, since GMM-UBM uses a common anti-model, namely UBM, for all target speakers, it tends to be weak in rejecting impostors' voices that are similar to the target speaker's voice. To overcome this limitation, we propose a discriminative feedback adaptation (DFA) framework that reinforces the discriminability between the target speaker model and the antimodel, while preserves the generalization ability of the GMM-UBM approach. This is done by adapting the UBM to a target-speakerdependent anti-model based on a minimum verification squarederror criterion, rather than estimating from scratch by applying the conventional discriminative training schemes. The results of experiments conducted on the NIST2001-SRE database show that DFA substantially improves the performance of the conventional GMM-UBM approach.Index Terms-Discriminative feedback adaptation, loglikelihood ratio, minimum verification squared-error linear regression, speaker verification

show abstract

Section: Minimum Verification Squared-error (Mvse) Adaptation Strategymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Discriminative Feedback Adaptation for GMM-UBM Speaker Verification

Chao

Tsai

Wang

2008

2008 6th International Symposium on Chinese Spoken Language Processing

View full text Add to dashboard Cite

show abstract

“…To obtain an optimal set of weights, we propose using Minimum Verification Error (MVE) training [6,7].…”

Section: Minimum Verification Error Trainingmentioning

confidence: 99%

“…In contrast to the geometric mean L 3 (U) or the arithmetic mean L 1 (U), which are independent of the system training, our combination scheme treats the background models unequally according to how close each individual is to the hypothesized speaker model, and quantifies the unequal nature of the background models by a set of weights optimized in the training phase. The optimization is carried out by Minimum Verification Error (MVE) training [6,7], which minimizes both the false acceptance probability and the false rejection probability.…”

Section: Introductionmentioning

confidence: 99%

Improved Methods for Characterizing the Alternative Hypothesis using Minimum Verification Error Training for LLR-Based Speaker Verification

Chao

Tsai

Wang

et al. 2007

2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07

View full text Add to dashboard Cite

Speaker verification based on the log-likelihood ratio (LLR) is essentially a task of modeling and testing two hypotheses: the null hypothesis and the alternative hypothesis. Since the alternative hypothesis involves unknown imposters, it is usually hard to characterize a priori. In this paper, we propose a framework to better characterize the alternative hypothesis with the goal of optimally separating client speakers from imposters. The proposed framework is built on either a weighted arithmetic combination or a weighted geometric combination of useful information extracted from a set of pre-trained anti-speaker models. The parameters associated with the combinations are then optimized using Minimum Verification Error training such that both the false acceptance probability and the false rejection probability are minimized. Our experiment results show that the proposed framework outperforms conventional LLR-based approaches.

show abstract

“…[4] proposed the use of general Gaussian mixture models which offer improved speech modeling resulting in better verification accuracy; Ref. [5] reported significant performance improvement using minimum verification error training; and Li et al in [6] proposed the method of utterance verification embedded in a human-machine dialog which can be used for both automatic registration and speaker verification. In this work, we focus on the issue of speech coding and its impact on the performance of a speaker verification system due to the fact that nearly all telecommunication networks today are digital and thus speech signals that are being transmitted through the networks are all encoded into bit-streams at various bit rates.…”

Section: Introductionmentioning

confidence: 99%

Speaker Verification Using Coded Speech

Moreno-Daniel

Juang

Nolazco-Flores

2004

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The implementation of a pseudo text-independent Speaker Verification system is described. This system was designed to use only information extracted directly from the coded parameters embedded in the ITU-T G.729 bitstream. Experiments were performed over the YOHO database [1]. The feature vector as a short-time representation of speech consists of 16 LPC-Cepstral coefficients, as well as residual information appended in the form of a pitch estimate and a measure of vocality of the speech. The robustness in verification accuracy is also studied. The results show that while speech coders, G.729 in particular, introduce coding distortions that lead to verification performance degradation, proper augmented use of unconventional information nevertheless leads to a competitive performance on par with that of a well-studied traditional system which does not involve signal coding and transmission. The result suggests that speaker verification over a cell phone connection remains feasible even though the signal has been encoded to 8 Kb/s.

show abstract

Speaker verification using minimum verification error training

Cited by 30 publications

References 7 publications

Discriminative Feedback Adaptation for GMM-UBM Speaker Verification

Discriminative Feedback Adaptation for GMM-UBM Speaker Verification

Improved Methods for Characterizing the Alternative Hypothesis using Minimum Verification Error Training for LLR-Based Speaker Verification

Speaker Verification Using Coded Speech

Contact Info

Product

Resources

About