A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf
.
Filters
Speaker Identification Using Data-Driven Score Classification
2016
Image Processing & Communications
We suggest the use of fusion with a discriminative algorithm such as a Support Vector Machine in a real-world speaker identification application where the text-independent scenario predominates based on ...
We present a comparative evaluation of different classification algorithms for a fusion engine that is used in a speaker identity selection task. ...
Speaker Identity Selection The traditional speaker identification decision is based on the selection of the maximum score, i.e. the speaker model with the maximum likelihood to have produced the input ...
doi:10.1515/ipc-2016-0011
fatcat:sdmblym6pjetvhyiu6vno5pmhu
Enhanced VQ-Based Algorithms for Speech Independent Speaker Identification
[chapter]
2003
Lecture Notes in Computer Science
Weighted distance measure and discriminative training are two different approaches to enhance VQ-based solutions for speaker identification. ...
Then it proposes two new algorithms combining the heuristic weighting and the partition normalized distance measure with group vector quantization discriminative training to take advantage of both approaches ...
Testing Results A new approach combining the weighted distance measure and the discriminative training is proposed to enhance VQ-based solutions for speech independent speaker identification. ...
doi:10.1007/3-540-44887-x_56
fatcat:hgdcho7wbzd6pfuzziqwk2rixm
Speaker Identification Using Discriminative Learning of Large Margin GMM
[chapter]
2011
Lecture Notes in Computer Science
In an earlier work, we proposed an algorithm for discriminative training of GMM with diagonal covariances under a large margin criterion. ...
They are generally trained using the generative criterion of maximum likelihood estimation. ...
For two utterances x and y, a kernel distance based on the Kullback-Leibler divergence between the GMM models trained on these utterances [4] , is defined as: K(x, y) = M m=1 √ w m Σ −(1/2) m µ xm T √ ...
doi:10.1007/978-3-642-24958-7_35
fatcat:3yuudg2vrjbujbhv4rdxovj3yy
A Kullback-Leibler Divergence Based Kernel for SVM Classification in Multimedia Applications
2003
Neural Information Processing Systems
We derive a kernel distance based on the Kullback-Leibler (KL) divergence between generative models. ...
Fisher kernel based SVM's and the generative classifiers in speaker identification and image classification. ...
The underlying generative model was the same one used for the GMM classification experiments. The task of speaker verification is different from speaker identification. ...
dblp:conf/nips/MorenoHV03
fatcat:tafniovbhjfj7gpq5mx6gd3kke
Weight Based Super-Gmm For Speaker Identification Systems
2008
Zenodo
The EM algorithm provides maximum-likelihood (ML) estimates for the unknown model parameter from a training database. ...
In general as the number of speakers enrolled in the system increases, the separability (interspeaker distance) between speaker models decreases, thereby causing a degradation on the identification/ recognition ...
doi:10.5281/zenodo.41235
fatcat:ajgcfcx5jbaovjrorvla7cajpe
Sparse Representation for Speaker Identification
2010
2010 20th International Conference on Pattern Recognition
Experiments have been conducted on the standard TIMIT [14] database and a comparison with the state-of-art speaker identification algorithms yields a favorable performance index for the proposed algorithm ...
We address the closed-set problem of speaker identification by presenting a novel sparse representation classification algorithm. ...
Motivated with these studies, we propose a novel speaker identification algorithm based on sparse representation. ...
doi:10.1109/icpr.2010.1083
dblp:conf/icpr/NaseemTB10a
fatcat:tfcaax7zhravlav6mxph5iw7ji
Automatic Language Identification with Discriminative Language Characterization Based on SVM
2008
IEICE transactions on information and systems
In this paper, in order to increase the dimension of the score vector and alleviate the inter-speaker variability within the same language, multiple data groups based on supervised speaker clustering are ...
Besides, the performances of proposed PPRLM and GMMs algorithms achieve an EER of 5.1% and 5.0% respectively. key words: language identification, supervised speaker clustering, support vector machine, ...
The baseline system is built based on maximum likelihood (ML) algorithm, and each Gaussian mixture model corresponds to one target language. ...
doi:10.1093/ietisy/e91-d.3.567
fatcat:g4svdz3mn5c5nowgaerlmrnpyy
Discriminative training of Gaussian mixture speaker models: A new approach
2010
2010 National Conference On Communications (NCC)
Conventional speaker recognition systems use Gaussian mixture models (GMM) to model a speaker's voice based on the speaker's acoustic characteristics. ...
To increase the discriminative properties of a GMM for each speaker, a new approach that includes both positive and negative examples during the speaker training process is proposed. ...
[6] and [7] , discuss a Maximum Model Distance algorithm for HMMs (Hidden Markov Models) that has been extended to GMMs in [8] . ...
doi:10.1109/ncc.2010.5430204
fatcat:4ktqh4cparfurhip6d4zudwucy
Class-Discriminative Weighted Distortion Measure for VQ-based Speaker Identification
[chapter]
2002
Lecture Notes in Computer Science
We consider the distortion measure in vector quantization based speaker identification system. ...
The question how to benefit from the different discrimination power of phonemes in VQ-based speaker recognition returns into question how to assign discriminative weights for different code vectors and ...
The first one is to improve separability in the training phase by discriminative training algorithms. Examples in the VQ context are LVQ [12] and GVQ [8] algorithms. ...
doi:10.1007/3-540-70659-3_71
fatcat:pj7j2r6x4rdvfhhorwtpmpjyqm
Discriminative Analysis of Lip Motion Features for Speaker Identification and Speech-Reading
2006
IEEE Transactions on Image Processing
The best lip motion features for speaker identification are considered to be those that result in the highest discrimination of individual speakers in a population, whereas for speech-reading, the best ...
Index Terms-Bayesian discriminative feature selection, lip motion, speaker identification, speech recognition, temporal discriminative feature selection. ...
Coulon for their valuable help in implementation of the lip tracking module in our recognition system. ...
doi:10.1109/tip.2006.877528
pmid:17022256
fatcat:5py73decinhbdjcjiw5ra4laoe
Automatic Attendance System Using Speaker Recognition
2018
International Journal of Trend in Scientific Research and Development
The system uses text dependent open identification with MFCC features and vector quantization based speaker modeling for authenticating the user. ...
A simple Euclidean distance scoring is used as the classifier. For decision making, the new approach, mean value threshold is proposed to optimize the system performance. ...
The system uses text dependent open-set speaker identification with MFCC features and vector quantization based speaker modeling for authenticating the user. ...
doi:10.31142/ijtsrd18763
fatcat:eblm42pv4faitlpf7pzckrggpu
Speaker verification using large margin GMM discriminative training
2011
2011 International Conference on Multimedia Computing and Systems
Index Terms-Large margin training, Gaussian mixture models, discriminative learning, speaker recognition, speaker verification ...
In an earlier work, we proposed an algorithm for discriminative training of GMM with diagonal covariances under a large margin criterion. ...
INTRODUCTION Most of state-of-the-art speaker recognition systems rely on the generative training of Gaussian Mixture Models (GMM) using maximum likelihood estimation and maximum a posteriori estimation ...
doi:10.1109/icmcs.2011.5945650
fatcat:av354krzh5f6lcm5efhvqhiiee
Using SVM as Back-End Classifier for Language Identification
2008
EURASIP Journal on Audio, Speech, and Music Processing
Robust automatic language identification (LID) is a task of identifying the language from a short utterance spoken by an unknown speaker. ...
One of the mainstream approaches named parallel phone recognition language modeling (PPRLM) has achieved a very good performance. ...
feed-forward multilayer neural network training, many algorithms are based on the gradient descent algorithms, such as back propagation (BP). ...
doi:10.1155/2008/674859
fatcat:rgw5u37nyfadlgledisvy3j76u
The question of disguised voice
2008
Journal of the Acoustical Society of America
The choice of the disguises is based on the most common ones used by offenders. ...
The purpose of this paper is to present the applications of statistical algorithms in order to detect and identify four specific disguises. ...
Identification based on a GMM classifier The idea of this method is to build a specific model for each kind of disguise based on GMM. ...
doi:10.1121/1.2935782
fatcat:2jrshm72b5acfmpvh5vegyitne
Discriminative speaker recognition using large margin GMM
2012
Neural computing & applications (Print)
Most state-of-the-art speaker recognition systems are based on discriminative learning approaches. ...
In an earlier work, we proposed an algorithm for discriminative training of GMM with diagonal covariances under a large margin criterion. ...
Acknowledgements The authors would like to thank the anonymous reviewers for their helpful comments. ...
doi:10.1007/s00521-012-1079-y
fatcat:3axaz7mxxvd5lbtzgydhqs5pua
« Previous
Showing results 1 — 15 out of 7,774 results