SUN’IY INTELLEKT ASOSIDA SO‘ZLOVCHINI AUTENTIFIKATSIYA QILISH TIZIMINI ISHLAB CHIQISH VA BAHOLASH
Keywords:
Nutq signali, nutq signallariga ishlov berish, so‘zlovchilarni aniqlash, so‘zlovchilarni tekshirish, so‘zlovchilarni autentifikatsiya qilish, filtrlash, log-Mel energiya, ECAPA-TDNNAbstract
Ushbu maqolada chuqur neyron tarmoqlarga asoslangan sun’iy intellekt tizimi yordamida inson ovozi orqali foydalanuvchini ishonchli tarzda aniqlovchi va tekshiruvchi so‘zlovchini autentifkatsiya qilish tizimi ishlab chiqish usuli va algoritmlari tahlil qilindi. Taklif etilayotgan tizimda signalga dastlabki ishlov berish, akustik xususiyatlarni ajratish va neyron tarmoq yordamida so‘zlovchilarni ifodalovchi xususiyatlar vektorlarini shakllantirish jarayonlarini integratsiyalashgan tizimda birlashtiradi. Dastlabki ishlov berish jarayonlarida signallarni filtrlash algoritmlari bilan filtrlanadi va signal normallashtiriladi. Xususiyatlarni ajratish bosqichida log-Mel energiya koeffisentlari ajratib olinadi va uni ECAPA-TDNN modeli uchun kirish parametrlari sifatida olinadi. ECAPA-TDNN modelidan so‘zlovchini ifodalovchi xususiyatlar vektori shakllantiriladi. Oxirgi bosqichda solishtirish algoritmlari yordamida xususiyatlar vektorlari solishtiriladi. Agar so‘zlovchilar xususiyatlari mos bo‘lsa tizim kirish uchun ruxsat beradi.
References
1. Abrorovich, Q. A. L., O‘G‘Li, X. S. I., & O‘G‘Li, X. U. K. (2024). SO‘ZLOVCHINI TANIB OLISHDA TANIB OLISH USULLARI TAHLILI. Eurasian Journal of Technology and Innovation, 2(6), 60-67.
2. Shukurov, K., Khasanov, U., Turaev, B., & Kakhkharov, A. L. (2023). The Effectiveness of the Implementation of Speech Command Recognition Algorithms in Embedded Systems. The Eurasia Proceedings of Science Technology Engineering and Mathematics, 23, 220-224.
3. Elbobo o‘g‘li, S. K., & Komiljon o‘g‘li, X. U. (2025, February). TIBBIY DIAGNOSTIKADA NUTQ SIGNALLARINING PARALINGVISTIK XUSUSIYATLARINI QO ‘LLASH. In INTERNATIONAL SCIENTIFIC RESEARCH CONFERENCE (Vol. 3, No. 31, pp. 118-121).
4. Kozlov, Alexander & Kudashev, Oleg & Matveev, Yu & Pekhovsky, Timur & Simonchik, Konstantin & Shulipa, Andrey. (2013). Speaker recognition system for the NIST SRE 2012. SPIIRAS Proceedings. 25. 350-370. 10.15622/sp.25.18.
5. Shukurov, K. (2021). Analysis of algorithms and implementation of real time speaker identification system. Bulletin of TUIT: Management and Communication Technologies, 4(4), 2.
6. Grozdić, Đorđe & Jovičić, Slobodan & Saric, Zoran & Subotić, Irina. (2015). Comparison of GMM/UBM and i-vector based speaker recognition systems
7. Kanagasundaram, Ahilan. (2014). Speaker Verification using I-vector Features. 10.13140/2.1.2933.7283.
8. Xu, Longting & Das, Rohan & Yılmaz, Emre & Yang, Jichen & Li, Haizhou. (2018). Generative x-vectors for text-independent speaker verification. 10.48550/arXiv.1809.06798.
9. Dawalatabad, Nauman & Ravanelli, Mirco & Grondin, François & Thienpondt, Jenthe & Desplanques, Brecht & Na, Hwidong. (2021). ECAPA-TDNN Embeddings for Speaker Diarization. 3560-3564. 10.21437/Interspeech.2021-941.
