Article 1318

Title of the article

EFFICIENCY ESTIMATING SOFTWARE FOR SPEECH RECOGNITION TECHNOLOGIES 

Authors

Alekseev Il'ya Vladimirovich, Postgraduate student, Penza State University (40 Krasnaya street, Penza, Russia), E-mail: aius@pnzgu.ru
Mitrokhin Maksim Aleksandrovich, Doctor of engineering sciences, head of sub-department of computer engineering, Penza State University (40 Krasnaya street, Penza, Russia), E-mail: vt@pnzgu.ru
Kol'chugina Elena Anatol'evna, Doctor of engineering sciences, professor, sub-department of mathematical support and computer application, Penza State University (40 Krasnaya street, Penza, Russia), E-mail: kea@pnzgu.ru 

Index UDK

004.9341 

DOI

10.21685/2072-3059-2018-3-1 

Abstract

Background. The object of research is modern technologies of speech recognition. The subject of the study is the evaluation of the effectiveness of modern speech recognition systems. The purpose of the work is to determine the main performance indicators of modern speech recognition technologies by the example of some systems to determine the possibility of their use in the speech interface of special purpose systems.
Materials and methods. The research was carried out using the methods of pattern recognition and methods of mathematical statistics.
Results. The efficiency of speech recognition systems was evaluated, quantitative indicators of accuracy and error probability were obtained in recognition of spoken control commands.
Conclusions. The existing systems of speech recognition of general purpose have quite high efficiency, but none of the systems considered cannot be used in special systems at this stage. 

Key words

user interface, speech interface, hidden markov model, neural networks, speech recognition 

 Download PDF
References

1. Kulikov S. S. Testirovanie programmnogo obespecheniya. Bazovyy kurs: prakt. posobie [Software testing. Basic practical course]. Minsk: Chetyre chetverti, 2015, 294 p.
2. GitHub. Available at: https://cmusphinx.github.io/wiki/tutorial/ (accessed Jul. 17, 2018).
3. Mozilla Research. Available at: https://research.mozilla.org/machine-learning/(accessed Jul. 17, 2018).
4. TensorFlow. Available at: https://www.tensorflow.org/ (accessed Jul. 17, 2018).
5. Yandex. Available at: https://tech.yandex.ru/speechkit/mobilesdk/doc/common/speechkit-common-asr-overview-technology-docpage/ (accessed Jul. 17, 2018).
6. Auditech. Available at: http://www.auditech.ru/page/darkness.html (accessed Jul. 17, 2018).

 

Дата создания: 19.04.2019 13:58
Дата обновления: 22.04.2019 07:55