journal article·2019-04-17Open access

Unsupervised Learning of Total Variability Embedding for Speaker Verification with Random Digit Strings

Woo Hyun Kang, Nam Soo Kim

Applied Sciences· Volume 9· Issue 8

Abstract

Recently, the increasing demand for voice-based authentication systems has encouraged researchers to investigate methods for verifying users with short randomized pass-phrases with constrained vocabulary. The conventional i-vector framework, which has been proven to be a state-of-the-art utterance-level feature extraction technique for speaker verification, is not considered to be an optimal method for this task since it is known to suffer from severe performance degradation when dealing with short-duration speech utterances. More recent approaches that implement deep-learning techniques for embedding the speaker variability in a non-linear fashion have shown impressive performance in various speaker verification tasks. However, since most of these techniques are trained in a supervised manner, which requires speaker labels for the training data, it is difficult to use them when a scarce amount of labeled data is available for training. In this paper, we propose a novel technique for extracting an i-vector-like feature based on the variational autoencoder (VAE), which is trained in an unsupervised manner to obtain a latent variable representing the variability within a Gaussian mixture model (GMM) distribution. The proposed framework is compared with the conventional i-vector method using the TIDIGITS dataset. Experimental results showed that the proposed method could cope with the performance deterioration caused by the short duration. Furthermore, the performance of the proposed approach improved significantly when applied in conjunction with the conventional i-vector framework.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: Speech Recognition and Synthesis · Speech and Audio Processing · Music and Audio Processing

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

Department of Electrical and Computer Engineering and the Institute of New Media and Communications, Seoul National University, Seoul 08826, Korea

Funding

Samsung Research Funding Center of Samsung Electronics

References (34)

Show source-provided references
1. Speaker recognition by machines and humans10.1109/MSP.2015.2462851 ↗
2. Support vector machines using GMM supervectors for speaker verification10.1109/LSP.2006.870086 ↗
3. Eigenvoice modeling with sparse training data10.1109/TSA.2004.840940 ↗
4. Dehak, N., Kenny, P., Dehak, R., Glembek, O., Dumouchel, P., Burget, L., Hubeika, V., and Castaldo, F. (2009, January 19–24). Support vector machines and joint factor analysis for speaker verification. Proceedings of the 2009 IEEE International Conference on Acoustics, Speech and Signal Processing, Taipei, Taiwan.10.1109/ICASSP.2009.4960564 ↗
5. Front-end factor analysis for speaker verification10.1109/TASL.2010.2064307 ↗
6. Kenny, P. (2012, January 25–28). A small footprint i-vector extractor. Proceedings of the Odyssey, Singapore.
7. Variani, E., Lei, X., McDermott, E., Moreno, I.L., and Gonzalez-Dominguez, J. (2014, January 4–9). Deep neural networks for small footprint text-dependent speaker verification. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, Florence, Italy.10.1109/ICASSP.2014.6854363 ↗
8. Snyder, D., Garcia-Romero, D., Povey, D., and Khudanpur, S. (2017, January 20–24). Deep neural network embeddings for text-independent speaker verification. Proceedings of the INTERSPEECH, Stockholm, Sweden.10.21437/Interspeech.2017-620 ↗
9. Snyder, D., Garcia-Romero, D., Sell, G., Povey, D., and Khudanpur, S. (2018, January 22–27). X-vectors: Robust DNN embeddings for speaker recognition. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, Prague, Czech Republic.10.1109/ICASSP.2018.8461375 ↗
10. Tang, Y., Ding, G., Huang, J., He, X., and Zhou, B. (2019, January 12–17). Deep speaker embedding learning with multi-level pooling for text-independent speaker verification. Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing, Brighton, UK.10.1109/ICASSP.2019.8682712 ↗
11. Fang, X., Zou, L., Li, J., Sun, L., and Ling, Z.H. (2019, January 12–17). Channel adversarial training for cross-channel text-independent speaker recognition. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, Brighton, UK.10.1109/ICASSP.2019.8682327 ↗
12. Chowdhury, F.A.R.R., Wang, Q., Moreno, I.L., and Wan, L. (2018, January 22–27). Attention-based models for text-dependent speaker verification. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, Prague, Czech Republic.
13. Wan, L., Wang, Q., Papir, A., and Moreno, I.L. (2018, January 22–27). Generalized end-to-end loss for speaker verification. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, Prague, Czech Republic.10.1109/ICASSP.2018.8462665 ↗
14. Speaker-phonetic i-vector modeling for text-dependent speaker verification with random digit strings10.1587/transinf.2018EDP7310 ↗
15. Saeidi, R., and Alku, P. (2015, January 6–10). Accounting for uncertainty of i-vectors in speaker recognition using uncertainty propagation and modified imputation. Proceedings of the INTERSPEECH, Dresden, Germany.10.21437/Interspeech.2015-703 ↗
16. Hasan, T., Saeidi, R., Hansen, J.H.L., and van Leeuwen, D.A. (2013, January 26–31). Duration mismatch compensation for i-vector based speaker recognition systems. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, Vancover, ON, Canada.10.1109/ICASSP.2013.6639154 ↗
17. Quality measure functions for calibration of speaker recognition systems in various duration conditions10.1109/TASL.2013.2279332 ↗
18. Mandasari, M.I., McLaren, M., and van Leeuwen, D.A. (2011, January 28–31). Evaluation of i-vector speaker recognition systems for forensic application. Proceedings of the INTERSPEECH, Florence, Italy.10.21437/Interspeech.2011-6 ↗
19. Kingma, D.P., and Welling, M. (2013). Auto-encoding variational Bayes. arXiv.
20. Doersch, C. (2016). Tutorial on variational autoencoders. arXiv.
21. Park, J., Kim, D., Lee, J., Kum, S., and Nam, J. (2018, January 14). A Hybrid of deep audio feature and i-vector for artist recognition. Proceedings of the Joint Workshop on Machine Learning for Music, International Conference on Machine Learning (ICML), Stockholm, Sweden.
22. Speaker verification using adapted Gaussian mixture models10.1006/dspr.1999.0361 ↗
23. Dehak, N., Torres-Carrasquillo, P.A., Reynolds, D., and Dehak, R. (2011, January 28–31). Language recognition via ivectors and dimensionality reduction. Proceedings of the INTERSPEECH, Florence, Italy.10.21437/Interspeech.2011-328 ↗
24. Learning deep generative models10.1146/annurev-statistics-010814-020120 ↗
25. Leonard, R.G. (1984, January 19–21). A database for speaker-independent digit recognition. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, San Diego, CA, USA.
26. Phone recognition on the TIMIT database
27. Gravier, G. (2019, March 01). SPro: Speech Signal Processing Toolkit. Available online: http://gforge.inria.fr/projects/spro.
28. Sadjadi, S.O., Slaney, M., and Heck, L. (2013, January 21–22). MSR identity toolbox v1.0: A MATLAB toolbox for speaker recognition research. Proceedings of the Speech and Language Processing Technical Committee Newsletter, Grenoble, France.
29. Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., Corrado, G.S., Davis, A., Dean, J., and Devin, M. (2019, March 01). Tensorflow: Large-Scale Machine Learning Heterogenous Systems. Available online: tensorflow.org.
30. Adaptive subgradient methods for online learning and stochastic optimization
31. Dropout: A simple way to prevent neural networks from overfitting
32. Garcia-Romero, D., and Espy-Wilson, C.Y. (2011, January 28–31). Analysis of i-vector length normalization in speaker recognition systems. Proceedings of the INTERSPEECH, Florence, Italy.10.21437/Interspeech.2011-53 ↗
33. (2019, March 01). The NIST Year 2008 Speaker Recognition Evaluation Plan, Available online: http://www.itl.nist.gov/iad/mig//tests/sre/2008/.
34. (2019, March 01). The NIST Year 2010 Speaker Recognition Evaluation Plan, Available online: http://www.itl.nist.gov/iad/mig//tests/sre/2010/.