A Study of Interspeaker Variability in Speaker Verification

By Patrick Kenny, Pierre Ouellet, Najim Dehak, Vishwa Gupta, Pierre Dumouchel. IEEE Transactions on Audio, Speech, and Language Processing, 2008

704 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/eaad1dc7142cf33fef94438c931bd40741a2e74a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Signal Processing

Citations in the corpus

Support vector machines versus fast scoring in the low-dimensional total variability space for speaker verification (2009, methodology): “In [1], Kenny et al. described how the speaker-dependent supervector and channel-dependent supervector can be represented in low-dimensional spaces.”
Spoofing and countermeasures for speaker verification: A survey (2015, methodology): “Parallel to the development of SVM-based discriminative approaches, generative factor analysis models were pioneered in Kenny (2006), Kenny et al. (2007) and Kenny et al. (2008).”
Frame-by-frame language identification in short utterances using deep neural networks (2015, methodology): “point estimate of the latent variables in a factor analysis model [12] [13].”
Learning Speaker-Specific Characteristics With a Deep Neural Architecture (2011, methodology): “37 • JFA (Joint Factor Analysis): A system based on joint factor analysis [117] that uses GMM-UBM supervector approach [109], [114].”
A novel scheme for speaker recognition using a phonetically-aware deep neural network (2014, methodology): “The reader can refer to [11] for more details.”
Bayesian Speaker Verification with Heavy-Tailed Priors (2010, methodology): “A speaker verification system using these features and a simple classifier produced better results on the 2008 NIST speaker recognition evaluation (SRE) data than Joint Factor Analysis (JFA) [2, 3].”
Front-End Factor Analysis for Speaker Verification (2011, methodology): “Similar to [ 3 ], the Baum‐Welch statistics are extracted using the UBM.”
Language Recognition via i-vectors and Dimensionality Reduction (2011, methodology): “Each vector is then converted into a 56-dimensional vector following a shifted delta cepstral parameterization using a 7-1-3-7 scheme and concatenation to the static cepstral coefficients.”
Deep neural networks for small footprint text-dependent speaker verification (2014, methodology): “Background models vary from simple Gaussian mixture model (GMM) based Universal Background Models (UBMs) [1] to more sophisticated Joint Factor Analysis (JFA) based models [2, 3, 4].”
An overview of text-independent speaker recognition: From features to supervectors (2010, background): “An important recent advance in speaker recognition has been the development of explicit inter-session variability compensation techniques (Burget et al., 2007; Kenny et al., 2008; Vogt and Sridharan, 2008).”
A review of depression and suicide risk assessment using speech analysis (2015, background): “The negative effect of varying channel characteristics and techniques to mitigate the effects are widely studied in speaker recognition, (Kenny et al., 2008; Kinnunen and Li, 2009) and discussed further in relation to mental state in Section 8.”
Deep Speaker Embeddings for Short-Duration Speaker Verification (2017, methodology): “The GMM-UBM paradigm was followed successively by joint factor analysis (JFA) and the well-known i-vector speaker representation [3, 4].”
Speaker recognition based on deep learning: An overview (2020, background): “See [9, 30] for the details of training Tc and estimating the i-vectors.”
The subspace Gaussian mixture model - A structured model for speech recognition (2011, methodology): “This method has similarities to Joint Factor Analysis as used in speaker recognition (Kenny et al., 2008) and to Eigenvoices (Kuhn et al., 2001) and Cluster Adaptive Training (CAT) (Gales, 2001) as proposed for speech recognition.”
Text-dependent speaker recognition using PLDA with uncertainty propagation (2013, background): “Joint-Factor Analysis and its successor, the combination of i-vectors and Probabilistic Linear Discriminant Analysis (PLDA) have become the dominant approaches in the field, as the NIST and other contests can verify, [1], [2], [3].”
Speaker recognition by means of Deep Belief Networks (2013, methodology): “A Maximum-Likelihood estimate of matrix T is obtained by minor modifications of the Joint Factor Analysis approach [16][17].”
Full-covariance UBM and heavy-tailed PLDA in i-vector speaker verification (2011, methodology): “PLDA can be seen as a special case of Joint Factor Analysis (JFA) [2] with a single Gaussian component.”
Comparison of Speaker Recognition Approaches for Real Applications (2011, methodology): “: Investigation into variants of Joint Factor Analysis for speaker recognition, In: Proc.”