A Noise-Robust System for NIST 2012 Speaker Recognition Evaluation

Ferrer, Luciana; McLaren, Mitchell; Scheffer, Nicolas; Lei, Yun; Graciarena, Martin; Mitra, Vikramjit

doi:10.21236/ada614010

Cited by 20 publications

(24 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If xt is the extracted speaker factor and if m is the supervector of an UBM of the acoustic feature vectors encountered in the training data, then the GMM supervector mt representing a window centered around the considered frame is approximated by mt = m + V xt (4) with V being the eigenvoice matrix. The speaker factor extraction at time t considers the frames inside a window of length Te centered around t. The procedure for extracting the speaker factors is similar to the iVector extraction described in [10].…”

Section: Boundary Generation With Speaker Factor Extractionmentioning

confidence: 99%

See 1 more Smart Citation

Soft VAD in Factor Analysis Based Speaker Segmentation of Broadcast News

Desplanques¹,

Demuynck²,

Martens³

2016

The Speaker and Language Recognition Workshop (Odyssey 2016)

View full text Add to dashboard Cite

In this work we propose to integrate a soft voice activity detection (VAD) module in an iVector-based speaker segmentation system. As speaker change detection should be based on speaker information only, we want it to disregard the nonspeech frames by applying speech posteriors during the estimation of the Baum-Welch statistics. The speaker segmentation relies on speaker factors which are extracted on a frame-byframe basis using an eigenvoice matrix. Speaker boundaries are inserted at positions where the distance between the speaker factors at both sides is large. A Mahalanobis distance seems capable of suppressing the effects of differences in the phonetic content at both sides, and therefore, to generate more accurate speaker boundaries. This iVector-based segmentation significantly outperforms Bayesian Information Criterion (BIC) segmentation methods and can be made adaptive on a file-by-file basis in a two-pass approach. Experiments on the COST278 multilingual broadcast news database show significant reductions of the boundary detection error rate by integrating the soft VAD. Furthermore, the more accurate boundaries induce a slight improvement of the iVector Probabilistic Linear Discriminant Analysis system that is employed for speaker clustering.

show abstract

Section: Boundary Generation With Speaker Factor Extractionmentioning

confidence: 99%

“…Note that the introduction of a voice activity detection (VAD) has already become common practice in related fields such as speaker recognition and language recognition (see e.g. [4,5]). …”

Section: Introductionmentioning

confidence: 99%

Soft VAD in Factor Analysis Based Speaker Segmentation of Broadcast News

Desplanques¹,

Demuynck²,

Martens³

2016

The Speaker and Language Recognition Workshop (Odyssey 2016)

View full text Add to dashboard Cite

show abstract

“…Even though one site continued to submit successful high-and low-level combined systems in those big data evaluations (Ferrer et al, 2013;Kajarekar et al, 2009;Scheffer et al, 2011), there was a consensus in turning back to cepstral-only systems. The computational complexity of higher-level systems and the relative improvements obtained in limited training data conditions helped the community to move towards a scientifically complex but very rewarding approach because of the performance and computational efficiency of new high-dimensional spectral systems as JFA-compensated GMM-UBM, and later, i-vector front-end extraction and PLDA based classification.…”

Section: Factor Analysis and I-vectorsmentioning

confidence: 99%

Evaluating Automatic Speaker Recognition systems: An overview of the NIST Speaker Recognition Evaluations (1996-2014)

González-Rodríguez¹

2014

loquens

View full text Add to dashboard Cite

Automatic Speaker Recognition systems show interesting properties, such as speed of processing or repeatability of results, in contrast to speaker recognition by humans. But they will be usable just if they are reliable. Testability, or the ability to extensively evaluate the goodness of the speaker detector decisions, becomes then critical. In the last 20 years, the US National Institute of Standards and Technology (NIST) has organized, providing the proper speech data and evaluation protocols, a series of text-independent Speaker Recognition Evaluations (SRE). Those evaluations have become not just a periodical benchmark test, but also a meeting point of a collaborative community of scientists that have been deeply involved in the cycle of evaluations, allowing tremendous progress in a specially complex task where the speaker information is spread across different information levels (acoustic, prosodic, linguistic…) and is strongly affected by speaker intrinsic and extrinsic variability factors. In this paper, we outline how the evaluations progressively challenged the technology including new speaking conditions and sources of variability, and how the scientific community gave answers to those demands. Finally, NIST SREs will be shown to be not free of inconveniences, and future challenges to speaker recognition assessment will also be discussed. NIST de reconocimiento de locutor (1996NIST de reconocimiento de locutor ( -2014.-Los sistemas automáticos de reconocimiento de locutor son críticos para la organización, etiquetado, gestión y toma de decisiones sobre grandes bases de datos de voces de diferentes locutores. Con el fin de procesar eficientemente tales cantidades de información de voz, necesitamos sistemas muy rápidos y, al no estar libre de errores, lo suficientemente fiables. Los sistemas actuales son órdenes de magnitud más rápidos que tiempo real, permitiendo tomar decisiones automáticas instantáneas sobre enormes cantidades de conversaciones. Pero tal vez la característica más interesante de un sistema automático es la posibilidad de ser analizado en detalle, ya que su rendimiento y fiabilidad puede ser evaluada de manera ciega sobre cantidades enormes de datos en una gran diversidad de condiciones. En los últimos 20 años, el Instituto Nacional de Estándares y Tecnología (NIST) de EE. UU. ha organizado, proporcionando los datos de voz y protocolos de evaluación adecuada, una serie de evaluaciones de reconocimiento de locutor independiente del texto. Esas evaluaciones se han convertido no sólo en una prueba comparativa periódica, sino también en punto de encuentro de una comunidad colaborativa de científicos que han estado profundamente involucrados en el ciclo de evaluaciones, lo que ha permitido un enorme progreso en una tarea especialmente compleja en la que la información individualizadora del locutor se encuentra dispersa en diferentes niveles de información (acústica, prosódica, lingüística...) y está fuertemente afectada por factores de variabilidad intrínsecos y extrínsecos al ...

show abstract

“…Both use simple GMM-based speech activity detection as defined in [12] and an i-vector/probabilistic linear discriminant analysis (PLDA) framework [13,14].…”

Section: Experimental Protocol and System Configurationmentioning

confidence: 99%

“…Full SRE'12 System: A gender-dependent system was trained based on the protocols used in the development of our SRE'12 submission [12] in order to evaluate the tuned features. The number UBM components increased to 2048 and was trained using a subset of 8000 clean speech samples; the i-vector subspace was trained using 51224 samples from which 600D i-vectors were extracted.…”

Section: Experimental Protocol and System Configurationmentioning

confidence: 99%

Effective use of DCTS for contextualizing features for speaker recognition

McLaren

Scheffer

Ferrer

et al. 2014

2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

This article proposes a new approach for contextualizing features for speaker recognition through the discrete cosine transform (DCT). Specifically, we apply a 2D-DCT transform on the Mel filterbank outputs to replace the common Mel frequency cepstral coefficients (MFCCs) appended by deltas and double deltas. A thorough comparison of algorithms for delta computation and DCT-based contextualization for speaker recognition is provided and the effect of varying the size of analysis window in each case is considered. Selection of 2D-DCT coefficients using a zig-zag approach permits definition of an arbitrary feature dimension using the most energized coefficients. We show that 60 coefficients computed using our approach outperforms the standard MFCCs appended with double deltas by up to 25% relative on the NIST 2012 speaker recognition evaluation (SRE) corpus in both Cprimary and equal error rate (EER) while additional coefficients increase system robustness to noise.

show abstract

A Noise-Robust System for NIST 2012 Speaker Recognition Evaluation

Cited by 20 publications

References 5 publications

Soft VAD in Factor Analysis Based Speaker Segmentation of Broadcast News

Soft VAD in Factor Analysis Based Speaker Segmentation of Broadcast News

Evaluating Automatic Speaker Recognition systems: An overview of the NIST Speaker Recognition Evaluations (1996-2014)

Effective use of DCTS for contextualizing features for speaker recognition

Contact Info

Product

Resources

About