Automatic Speaker Verification

Statistics of frame-level pitch have recently been used in speaker recognition systems with good results [1, 2, 3]. Although they convey useful long-term information about a speaker's distribution of f 0 values, such statistics fail to capture information about local dynamics in intonation that characterize an individual's speaking style. In this work, we take a first step toward capturing such suprasegmental patterns for automatic speaker verification. Specifically, we model the speaker's f 0 movements by fitting a piecewise linear model to the f 0 track to obtain a stylized f 0 contour. Parameters of the model are then used as statistical features for speaker verification. We report results on 1998 NIST speaker verification evaluation. Prosody modeling improves the verification performance of a cepstrum-based Gaussian mixture model system (as measured by a task-specific Bayes risk) by 10%.

For a few years, the problem of session variability in text- independent automatic speaker verification is being tackled ac- tively. A new paradigm based on a factor analysis model have successfully been applied for this task. While very efficient, its implementation is demanding. In this paper, the algorithms in- volved in the eigenchannel MAP model are written down for a straightforward implementation, without referring to previous work or complex mathematics. In addition, a different com- pensation scheme is proposed where the standard GMM likeli- hood can be used without any modification to obtain good per- formance (even without the need of score normalization). The use of the compensated supervectors within a SVM classifier through a distance based kernel is also investigated. Experi- ments results shows an overall 50% relative gain over the stan- dard GMM-UBM system on NIST SRE 2005 and 2006 proto- cols (both at the DCFmin and EER). Index Terms: Speaker Verification, Session ...

Environmental noise and reverberation conditions severely degrade the performance of forensic speaker verification. Robust feature extraction plays an important role in improving forensic speaker verification performance. This paper investigates the effectiveness of combining features, mel frequency cepstral coefficients (MFCCs), and MFCC extracted from the discrete wavelet transform (DWT) of the speech, with and without feature warping for improving modern identity-vector (i-vector)-based speaker verification performance in the presence of noise and reverberation. The performance of i-vector speaker verification was evaluated using different feature extraction techniques: MFCC, feature-warped MFCC, DWT-MFCC, feature-warped DWT-MFCC, a fusion of DWT-MFCC and MFCC features, and fusion feature-warped DWT-MFCC and feature-warped MFCC features. We evaluated the performance of i-vector speaker verification using the Australian Forensic Voice Comparison and QUT-NOISE databases in the presence of noise, reverberation, and noisy and reverberation conditions. Our results indicate that the fusion of feature-warped DWT-MFCC and feature-warped MFCC is superior to other feature extraction techniques in the presence of environmental noise under the majority of signal-to-noise ratios (SNRs), reverberation, and noisy and reverberation conditions. At 0-dB SNR, the performance of the fusion of feature-warped DWT-MFCC and feature-warped MFCC approach achieves a reduction in average equal error rate of 21.33%, 20.00%, and 13.28% over feature-warped MFCC, respectively, in the presence of various types of environmental noises only, reverberation, and noisy and reverberation environments. The approach can be used for improving the performance of forensic speaker verification and it may be utilized for preparing legal evidence in court. INDEX TERMS Discrete wavelet transform, environmental noise and reverberation conditions, forensic speaker verification, feature warped-MFCC.

This article presents a new approach using the discrimination power of Support Vectors Machines (SVM) in combination with Gaussian Mixture Models (GMM) for Automatic Speaker Verification (ASV). In this combination SVMs are applied in the GMM model space. Each point of this space represents a GMM speaker model. The kernel which is used for the SVM allows the computation of a similarity between GMM models. It was calculated using the Kullback-Leibler (KL) divergence. The results of this new approach show a clear improvement compared to a simple GMM system on the NIST2005 Speaker Recognition Evaluation primary task.

a b s t r a c t Speaker verification has recently been introduced to the forensic field as a new and complimentary approach to other forensic methods. With the advancement in speech communication technologies including voice over IP and wireless multimedia applications, speech is seldom sent between two parties in plain, it is at least partially encrypted before transmission. We present automatic speaker verification techniques based on hidden Markov and Gaussian mixture models from partially encrypted speech from the perceptually less relevant speech features which are unencrypted. An equal error rate (EER) of 23% and minimum detection cost value of 8% has been achieved on a database of 84 speakers using adapted Gaussian mixture modeling. Comparison between different modeling techniques and effect of Gaussian mixture densities are also carried out and results are tabulated. The results suggest that partial or selective encryption techniques may provide content protection but will not protect the speaker's identity. ª (F. Iqbal). a v a i l a b l e a t w w w . s c i e n c e d i r e c t . c o m j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / d i i n d i g i t a l i n v e s t i g a t i o n 7 ( 2 0 1 0 ) 7 4 e8 0 1742-2876/$ e see front matter ª

A systematic analysis of user-dependent performance variability in the context of automatic speaker verification was first studied by . Different categories of users were distinguished and were called by animal names such as sheep, goats, lambs and wolves. Although such distinctions are important, it does not directly discriminate "well-behaved" users from "badly behaved" users. In our context, the badly behaved users are those who will bring the performance down when added to the system. We then extend such a study to formulate a user-specific score normalization (called F-norm's variant) and show that the user-dependent variability can be reduced to obtain an enhanced performance. By introducing some constraints, the proposed framework can also provide a stable user-dependent performance in terms of DET despite the fact that few (genuine) samples are available. In the context of multimodal biometrics, we show that it is possible to decide whether or not fusing the output of several systems is better than selecting any one of them, on a per user basis. This strategy is called an "OR-switcher". Based on 15 multimodal fusion experiments, the performance of OR-switcher is significantly better than the state-of-the-art score-level fusion algorithms.

This paper addresses the design and implementation of automatic speaker verification (ASV) systems. There is great interest in developing and increasing the performance of ASV applications, taking into account the advantages offered when ...

—Keeping track of the multiple passwords, PINs, memorable dates and other authentication details needed to gain remote access to accounts is one of modern life's less appealing challenges. The employment of a voice-based verification as a biometric technology for both children and adults could be a good replacement to the old fashioned memory dependent procedure. Using voice for authentication could be beneficial in several application areas, including, security, protection, education, call-based and web-based services. Voice-based biometric applications are subject to different types of spoofing attacks. The most accessible and affordable type of spoofing for a voice-based biometrics system is a replay attack. Replay, which is to playback a pre-recorded speech sample, presents a genuine risk to automatic speaker verification technology. This work presents two architectures for detecting frauds caused by replay attacks in a voice-based biometrics authentication systems. Experimental results confirmed that obtained performances from both methods could further improve by applying a machine learning algorithm for performing fusion at the score level. The performance of both methods further improved by fusion using independent sources of scores in different architectures.

Some preliminary investigations of within-speaker variations due to voluntary and induced speaking manners have been performed. The ultimate aim of the investigations was to suggest methods to take care of within-speaker variations in automatic speaker verification. Special software was developed to systematically elicit different types of voluntary and involuntary speech variations that might realistically occur in every-day situations. A database containing speech from 50 Swedish male speakers was collected using this software. Acoustic analyses have been performed on and the results compared between voluntary and involuntary speech variations. The acoustic parameters that have been studied included segment durations, formant frequencies at vowel midpoints, fundamental frequency and overall amplitude and amplitude in frequency bands.

It is well known that the speaker discriminative information is not equally distributed over the spectral domain. However, it is still not clear whether that distribution is altered when the speech is transmitted through telecommunication channels, which introduce different kinds of degradations. In this paper we address the analysis of different frequency sub-bands when the speech is distorted with different bandwidth filters and channel codecs, considering narrowband and wideband communications. Our i-vector experiments on different sub-bands with 782 speakers show that standard landline codecs perform generally better than wireless codecs due to their intrinsic coding algorithm, their performance being close to, but slightly worse than that of uncoded speech. Wideband signals offer significant benefits over narrowband for speaker verification. A smaller experiment with 21 speakers leads us to believe that the emerging super-wideband transmissions may provide even better results because it shows important speaker-specific content in the band 8-14kHz.

In this paper we report the experiment carried out on recently collected speaker recognition database namely Arunachali Language Speech Database (ALS-DB)to make a comparative study on the performance of acoustic and prosodic features for speaker verification task.The speech database consists of speech data recorded from 200 speakers with Arunachali languages of NorthEast India as mother tongue. The collected database is evaluated using Gaussian mixture model-Universal Background Model (GMM-UBM) based speaker verification system. The acoustic feature considered in the present study is Mel-Frequency Cepstral Coefficients (MFCC) along with its derivatives.The performance of the system has been evaluated for both acoustic feature and prosodic feature individually as well as in combination.It has been observed that acoustic feature, when considered individually, provide better performance compared to prosodic features. However, if prosodic features are combined with acoustic feature, performance of the system outperforms both the systems where the features are considered individually. There is a nearly 5% improvement in recognition accuracy with respect to the system where acoustic features are considered individually and nearly 20% improvement with respect to the system where only prosodic features are considered.

A Gaussian mixture model (GMM) is compared to a multiple binary classifier model (MBCM) in two speaker verification experiments conducted on telephone speech. The MBCM consists of 45 Moody-Darken radial basis function neural networks (MD-RBFNs) whose outputs are fused. Furthermore, the model is pruned in order to remove poorly performing MD-RBFNs. In the first experiment, true speakers and impostors are

Noise robustness of automatic speaker verification systems is crucial in real life applications. A study on the performance of several spectral subtraction-based speech enhancement techniques shows the poor performance of these algorithms when used as a preprocessing stage of the speaker verification system in the presence of colored noise. In this paper, a new technique based on the addition of modeled colored noise is introduced. Experimental results in both white and colored noise environments are presented, showing the improvement of the proposed scheme.

Alphanumeric passwords, identification via either chip cards or biometric parameters as iris pattern, fingerprints and voice are used to control access to confidential information and to restricted areas as well as to several types of automated money transactions. This paper aims at presenting an IP core whose purpose is to perform real-time speaker verification. The IP core can be used as part of a system to check if the speaker is really the one (he or she) who claims to be. For each speaker, the IP core needs a codebook which is obtained by off-line training. Linear Prediction Coding (LPC) and Vector Quantization (VQ) techniques are employed. The IP core was designed to process the signal in real time, meaning that, as soon as the IP core has gathered enough data to reach a certain confidence level, it will output its decision.