i-Vectors in speech processing applications: a survey

Verma, Pulkit; Das, Pradip K.

doi:10.1007/s10772-015-9295-3

Cited by 44 publications

(22 citation statements)

References 83 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…On the other hand, in the SITW database, 70% from FWMFCC features was fused with 30% from the corresponding CMVNPNCC features. The Future work will consider a similar extensive evaluation for a speaker identification system built from an I-vector approach [4].…”

Section: Discussionmentioning

confidence: 99%

“…According to [4], feature extraction within speaker identification should be less influenced by noise or the person's health. However, to improve the speaker identification accuracy (SIA), Mel frequency cepstral coefficients (MFCC) features were fused with inverse MFCC features (IMFCC) in [5], but the approach was limited by the number of GMM components.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of a speaker identification system with and without fusion using three databases in the presence of noise and handset effects

Al-Kaltakchi

Woo

Dlay

et al. 2017

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

In this study, a speaker identification system is considered consisting of a feature extraction stage which utilizes both power normalized cepstral coefficients (PNCCs) and Mel frequency cepstral coefficients (MFCC). Normalization is applied by employing cepstral mean and variance normalization (CMVN) and feature warping (FW), together with acoustic modeling using a Gaussian mixture model-universal background model (GMM-UBM). The main contributions are comprehensive evaluations of the effect of both additive white Gaussian noise (AWGN) and non-stationary noise (NSN) (with and without a G.712 type handset) upon identification performance. In particular, three NSN types with varying signal to noise ratios (SNRs) were tested corresponding to street traffic, a bus interior, and a crowded talking environment. The performance evaluation also considered the effect of late fusion techniques based on score fusion, namely, mean, maximum, and linear weighted sum fusion. The databases employed were TIMIT, SITW, and NIST 2008; and 120 speakers were selected from each database to yield 3600 speech utterances. As recommendations from the study, mean fusion is found to yield overall best performance in terms of speaker identification accuracy (SIA) with noisy speech, whereas linear weighted sum fusion is overall best for original database recordings.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Evaluation of a speaker identification system with and without fusion using three databases in the presence of noise and handset effects

Al-Kaltakchi

Woo

Dlay

et al. 2017

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

show abstract

“…Lowdimensional i-vector w is a dense representation of all relevant discrimination information. The backend of this classifier was a simplified GPLDA algorithm [7]. We used a log likelihood ratio of probabilities of two hypotheses.…”

Section: Classifiers Gaussian Mixture Models (Baseline Classifier)mentioning

confidence: 99%

“…The main idea behind our system is the fusion of scores [4] given by a set of binary classification algorithms to obtain better classification results. We opted for some common speaker modelling techniques such as Gaussian mixture models (GMM) [5], GMM based universal background models (UBM) and i-vectors [6,7]. The set of classifiers consists of multiple GMMs based classifier [1,5], i-vector-Gaussian Probabilistic Linear Discriminant Analysis (GPLDA) [7] and highly popular in machine learning community tree boosting method XGBoost [8].…”

Section: Introductionmentioning

confidence: 99%

“…We opted for some common speaker modelling techniques such as Gaussian mixture models (GMM) [5], GMM based universal background models (UBM) and i-vectors [6,7]. The set of classifiers consists of multiple GMMs based classifier [1,5], i-vector-Gaussian Probabilistic Linear Discriminant Analysis (GPLDA) [7] and highly popular in machine learning community tree boosting method XGBoost [8]. The most common choice of an acoustic feature for voice processing system is Mel frequency cepstral coefficients (MFCCs) [9].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Speech Replay Spoofing Attack Detection System Based on Fusion of Classification Algorithms

Lependin¹,

Филин²,

Malinin³

2018

Известия АлтГУ

View full text Add to dashboard Cite

Система обнаружения атак воспроизведением речи, основанная на смеси алгоритмов классификации А.А. Лепендин, Я.А. Филин, П.В. Малинин Алтайский государственный университет (Барнаул, Россия) Fast development of modern technologies of digital processing and speech recording leads to the fact that it is necessary to take into account the potential threats from the speech replay attacks. We propose our ensemble fusion replay attack detection system. It uses constant Q cepstral coefficients as speech features and short-time mean normalization for their preprocessing. The set of binary classifiers includes multiple Gaussian mixture models based Bayesian classifier, i-vector based Gaussian Probabilistic Linear Discriminant Analysis and XGBoost tree boosting algorithm. Fusion of scores was made by modified logistic regression algorithm from BOSARIS toolbox. ASV Spoof 2017 corpus is utilized in the experiments as the main database for anti-spoofing systems evaluation. Obtained results demonstrate that the proposed system can provide substantially better performance than the baseline Gaussian mixture model classifier. The pre-processing of cepstral features is crucial for the better performance of the system. High evaluation performance can be obtained using only few algorithms in a set. The attained value of equal error rate EER=12.44% for our fusion classifier is competitive with the best results obtained during last two years.Keywords: automatic speaker verification; voice spoofing; replay attacks; universal background model; i-vector; probabilistic linear discriminant analysis; tree boosting; model fusion.Быстрое развитие современных технологий циф-ровой обработки и записи речевых сигналов приве-ло к тому, что стал актуальным учет потенциальных угроз, связанных с атаками на биометрические си-стемы аутентификации, которые основаны на вос-произведении речи. В работе предложен подход к де-тектированию подобных атак при помощи ансамбля из нескольких классификаторов. В качестве инфор-мативных признаков речевого сигнала применя-лись Q-константные кепстральные коэффициенты. Проводилась их нормализация путем вычитания кепстрального среднего, оцениваемого на коротком временном интервале. Множество использованных бинарных классификаторов состояло из алгоритма на гауссовых смесях, гауссового вероятностного ли-нейного дискриминантного анализа в сочетании с из-влечением i-векторов речевых сигналов и алгоритма XGBoost. Смешивание осуществлялось при помощи модифицированного алгоритма логистической ре-грессии. Качество работы предложенного подхода оказалось существенно выше базового метода, осно-ванного на применении смесей гауссовых распреде-лений. Дополнительное существенное улучшение ка-чества было связано с предобработкой кепстральных коэффициентов. Было показано, что качество рабо-ты, близкое к наилучшему, может быть достигнуто при смешивании небольшого числа классификаторов. Достигнутое значение частоты ошибок EER = 12.44% для смеси классификаторов близко к лучшим из до-стигнутых к нынешнему моменту.Ключевые слова: верифик...

show abstract