Evaluation of Methods to Combine Different Speech Recognizers

• Autorzy: Tomas Rasymas , Vytautas Rudžionis
• Języki publikacji: EN

Abstrakty

EN

The paper deals with the problem of improving speech recognition by combining outputs of several different recognizers. We are presenting our results obtained by experimenting with different classification methods which are suitable to combine outputs of different speech recognizers. Methods which were evaluated are: k-Nearest neighbors (KNN), Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), Logistic Regression (LR) and maximum likelihood (ML). Results showed, that highest accuracy (98.16 %) was obtained when k-Nearest neighbors method was used with 15 nearest neighbors. In this case accuracy was increased by 7.78 % compared with best single recognizer result. In our experiments we tried to combine one native (Lithuanian language) and few foreign speech recognizers: Russian, English and two German recognizers. For the adaptation of foreign language speech recognizers we used text transcribing method which is based on formal rules. Our experiments proved, that recognition accuracy improves when few speech recognizers are combined.(original abstract)

Słowa kluczowe

EN

High-tech

PL

Nowe technologie

• Czasopismo: Annals of Computer Science and Information Systems
• Rocznik: 2015
• Tom: 5
• Strony: 1043--1047

Twórcy

autor

Tomas Rasymas

• Vilnius University

autor

Vytautas Rudžionis

• Vilnius University

Bibliografia

• R. Maskeliunas, A. Rudžionis, K. Ratkevičius, V. Rudžionis, "Investigation of Foreign Languages Models for Lithuanian Speech Recognition", Electronics and Electrical Engineering, no. 3(91), pp. 15- 20, 2009.
• V. Rudžionis, G. Raškinis, A. Rudžionis, K. Ratkevičius, "Comparative Analysis of Adapted Foreign Language and Native Lithuanian Speech Recognizers for Voice User Interface", Electronics and Electrical Engineering, vol. 19, no. 7, pp. 90-93, 2013.
• V. Rudžionis, G. Raškinis, A. Rudžionis, K. Ratkevičius, G. Bartišiute, "Web Services Based Hybrid Recognizer of Lithuanian Voice Commands", Electronics and Electrical Engineering, vol. 20, no. 9, pp. 50-53, 2014.
• T. Rasymas, V. Rudžionis, "Combining Multiple Foreign Language Speech Recognizers by using Neural Networks", Human Language Technologies - The Baltic Perspective, IOS Press, doi:10.3233/978-1- 61499-442-8-33, pp. 33-39, 2014.
• P. Kasparaitis, "Transcribing of the Lithuanian Text Using Formal Rules", Informatica, vol. 10, no. 4, pp. 367-376, 1999.
• P. Kasparaitis, "Lithuanian Speech Recognition Using the English Recognizer", Informatica, vol. 19, no. 4, pp. 505-516, 2008.
• V. Rudžionis, K. Ratkevičius, A. Rudžionis, G. Raškinis, R. Maskeliūnas, "Recognition of Voice Commands Using Hybrid Approach", ICIST2013, CCIS 403, Springer-Verlag Berlin, pp. 249-260, 2013.
• D. Huggins-Daines, M. Kumar, A. Chan, A. W Block, M. Ravishankar, A. I. Rudnicky, "Pocketsphinx: a free, real-time continuous speech recognition system for hand-held devices", IEEE ICASSP 2006 Proceedings, vol. 1, pp. 185-188, 2006.
• F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, J. Vanderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot, Duchesnay, "Scikit-learn: Machine Learning in Python", The Journal of Machine Learning Research, vol. 12, pp. 2825-2830, 2011.
• T. Schultz, A. Waibel, "Language-independent and language-adaptive acoustic modeling for speech recognition", Speech Communication 35 (1), 31-52, 2001.
• Z. Wang, T. Schultz, A. Waibel, "Comparison of Acoustic Model Adaptation Techniques on Non-Native Speech", IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), pp. 540-543, 2003.
• H. Meneido, J. Neto, "Combination of acoustic models in continuous speech recognition hybdrid systems", Proceedings of the International Conference in Spoken Language Processing, vol. 9, pp. 1000-1029, 2000.

Identyfikatory

DOI

10.15439/2015F62