To obtain access to full text of journal and articles you must register!
- Article name
- Analytical study of the problem of biometric identification and authentication of subjects by voice
- Authors
- Inivatov D. P., , daniilini@mail.ru, Omsk State Technical University, Omsk, Russia
- Keywords
- automatic speaker recognition / deep learning / spoofing / anti-noise / speaker voice parameters / biometric authentication / voice data corpora
- Year
- 2023 Issue 3 Pages 28 - 38
- Code EDN
- SEATMG
- Code DOI
- 10.52190/2073-2600_2023_3_28
- Abstract
- This article presents an overview of modern achievements and methods in the field of human voice recognition and biometric identification. The key challenges faced by researchers in this field, such as the fight against noise, voice drift, spoofing, diarization, contention attacks, etc. are considered. Comparative tables of the results of research conducted in this area are given. Particular attention is paid to the progress in the development of methods that can improve the accuracy of voice recognition under conditions of noise and variable voice characteristics. As a result of the review, the best values for the studied tasks are given: diarization - EER = 0.83%, noise control - 1.3%, emotion recognition - 8.95%. It is noted that a promising direction can be considered the implementation of various architectures, including a biometric-code neural network converter in order to protect biometric templates.
- Text
- BUY for read the full text of article
- Buy
- 500.00 rub
