Speaker diarization through speaker embeddings

This paper proposes to learn a set of high-level feature representations through deep learning, referred to as Speaker Embeddings, for speaker diarization. Speaker Embedding features are taken from the hidden layer neuron activations of Deep Neural Networks (DNN), when learned as classifiers to recognize a thousand speaker identities in a training set. Although learned through identification, speaker embeddings are shown to be effective for speaker verification in particular to recognize speakers unseen in the training set. In particular, this approach is applied to speaker diarization. Experiments, conducted on the corpus of French broadcast news ETAPE, show that this new speaker modeling technique decreases DER by 1.67 points (a relative improvement of about 8% DER).

[1]  Geoffrey E. Hinton,et al.  Phone Recognition with the Mean-Covariance Restricted Boltzmann Machine , 2010, NIPS.

[2]  Mickael Rouvier,et al.  A global optimization framework for speaker diarization , 2012, Odyssey.

[3]  Patrick Kenny,et al.  Support vector machines versus fast scoring in the low-dimensional total variability space for speaker verification , 2009, INTERSPEECH.

[4]  Simon Dobrisek,et al.  Incorporating Duration Information into I-Vector-Based Speaker Recognition Systems , 2014, Odyssey.

[5]  Olivier Galibert,et al.  The ETAPE corpus for the evaluation of speech-based TV content processing in the French language , 2012, LREC.

[6]  Yoshua Bengio,et al.  Word Representations: A Simple and General Method for Semi-Supervised Learning , 2010, ACL.

[7]  Yuning Jiang,et al.  Learning Deep Face Representation , 2014, ArXiv.

[8]  Patrick Kenny,et al.  Bayesian Speaker Verification with Heavy-Tailed Priors , 2010, Odyssey.

[9]  John H. L. Hansen,et al.  Duration mismatch compensation for i-vector based speaker recognition systems , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[10]  Mickael Rouvier,et al.  An open-source state-of-the-art toolbox for broadcast news diarization , 2013, INTERSPEECH.

[11]  Nicholas W. D. Evans,et al.  Short-Duration Speaker Modelling with Phone Adaptive Training , 2014, Odyssey.

[12]  Patrick Kenny,et al.  Front-End Factor Analysis for Speaker Verification , 2011, IEEE Transactions on Audio, Speech, and Language Processing.

[13]  Jean-François Bonastre,et al.  NON DIRECTLY ACOUSTIC PROCESS FOR COSTLESS SPEAKER RECOGNITION AND INDEXATION , 1999 .

[14]  Delphine Charlet,et al.  Speaker identification by location in an optimal space of anchor models , 2002, INTERSPEECH.

[15]  Daniel Garcia-Romero,et al.  Analysis of i-vector Length Normalization in Speaker Recognition Systems , 2011, INTERSPEECH.

[16]  Ieee Staff.  2017 25th European Signal Processing Conference (EUSIPCO) , 2017 .

[17]  Themos Stafylakis,et al.  PLDA for speaker verification with utterances of arbitrary duration , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[18]  Driss Matrouf,et al.  Variance-spectra based normalization for i-vector standard and probabilistic linear discriminant analysis , 2012, Odyssey.

[19]  Olivier Galibert,et al.  The First Official REPERE Evaluation , 2013, SLAM@INTERSPEECH.

[20]  James H. Elder,et al.  Probabilistic Linear Discriminant Analysis for Inferences About Identity , 2007, 2007 IEEE 11th International Conference on Computer Vision.

[21]  Daniel Povey,et al.  The Kaldi Speech Recognition Toolkit , 2011 .