Affiliation:
1. Nizhniy Novgorod State Linguistic University
2. National Research University Higher School of Economics
Abstract
We consider the task of automated quality control of sound recordings containing voice samples of individuals. It is shown that in this task the most acute is the small sample size. In order to overcome this problem, we propose the novel method of acoustic measurements based on relative stability of the pitch frequency within a voice sample of short duration. An example of its practical implementation using aninter-periodic accumulation of a speech signal is considered. An experimental study with specially developed software provides statistical estimates of the effectiveness of the proposed method in noisy environments. It is shown that this method rejects the audio recording as unsuitable for a voice biometric identification with a probability of 0,95 or more for a signal to noise ratio below 15 dB. The obtained results are intended for use in the development of new and modifying existing systems of collecting and automated quality control of biometric personal data. The article is intended for a wide range of specialists in the field of acoustic measurements and digital processing of speech signals, as well as for practitioners who organize the work of authorized organizations in preparing for registration samples of biometric personal data.
Publisher
FSUE VNIIMS All-Russian Research Institute of Metrological Service
Reference23 articles.
1. Savchenko V. V., Journal of Communications Technology and Electronics, 2016, vol. 61, no. 12, pp. 1374–1379. DOI: 10.1134/S1064226916120226
2. Kunesova M., Hruz M., Zajíc Z., Radova V., 21st International Conference on Speech
3. And Computer SPECOM-2019, Proceedings of the International Conference,
4. Istanbul, Turkey, August 20–25, 2019 Istanbul, Springer, 2019, pp. 247–257. DOI: 10.1007/978-3-030-26061-3_26
5. Joshi A., Kumar M., Das P. K, 2016-International Conference on Signal Processing and Communication (ICSC), Proceedings of the International Conference, Noida, India, IEEE Conference Publication, 2017, pp. 191–196. DOI: 10.1109/ICSPCom.2016.7980574
Cited by
4 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献