This article analyzes the effectiveness of using different models in speaker recognition processes and selects the best one for the system. In terms of accuracy and speed performance of the system, the classical MFCC + cosine similarity and modern x-vector, ECAPA-TDNN + PLDA architectures are compared. Based on the data set generated from different speakers, the accuracy, f1-score, EER, latency, and GPU load indicators of the models are evaluated. According to the experimental results, the ECAPA-TDNN model outperforms the other models with an accuracy of 95.7%. Since the speaker recognition stage is also important for speaker separation systems, accuracy indicators are of high relevance. The ECAPA-TDNN + PLDA model offers good solutions in terms of using computational resources, working with large data sets, and analyzing their data.
| Mualliflar | Шукуров, К.Э., Хасанов, У.К. |
|---|---|
| Jurnal | Рақамли технологияларнинг назарий ва амалий масалалари |
| Nashr sanasi | 2026-02-26 |
| Jild | 9 |
| Son | 1 |
| Betlar | 80-89 |
| Til | Rus |
| DOI | 10.62132/ijdt.v9i1.325 |
DOI: 10.62132/ijdt.v9i1.325 · Maqolaning asl sahifasi
идентификация говорящего, ECAPA-TDNN, x-вектор, MFCC, логарифмическое сходство, PLDA, косинусное сходство, глубокое обучение, вектор признаков, речевая биометрия, AM-softmax, speaker identification, ECAPA-TDNN, x-vector, MFCC, log-Mel, PLDA, cosine similarity, deep learning, feature vector, speech biometrics, AM-softmax
This article discusses the main issue under consideration, the most promising direction of data collection from sensors and data transmission using a wireless system, and technologies with a number of fundamentally new…
Today, the development of speech correction systems for general speech underdevelopment in preschool-age children using speech technologies is a pressing and pressing priority. This article describes the structure of a…
This article proposes a method and algorithm for solving a transportation problem in a fuzzy environment. This method is easy to understand and use, allowing for finding optimal solutions to real-world problems in fuzzy…
This article provides a comprehensive overview of the NEQR (Novel Enhanced Quantum Representation) model, designed for processing classical images on quantum computers. NEQR encodes image pixels directly as binary qubit…
This article examines the theoretical foundations of regulatory science in medicine as a discipline focused on the holistic control of living systems. Regulatory science investigates the mechanisms that coordinate…
This paper considers the problem of person re-identification (Re-ID), which plays a key role in intelligent video surveillance and behavior analysis systems. The main focus is on the comparative analysis of methods…
This paper investigates the hydrodynamic characteristics of dam-break flows using a three-dimensional numerical model based on the Reynolds-Averaged Navier–Stokes (RANS) equations and the Volume of Fluid (VOF) method…
The work proposes a mathematical model for monitoring and predicting the processes of heat and moisture exchange in a porous grain environment. The model represents the processes of heat transfer, moisture diffusion…
This paper proposes a fact-centric formal model and an incremental maintenance approach for a fact repository to represent heterogeneous digital information streams in a multi-agent environment. The approach covers fact…
This paper includes an extensive investigation of the magnetoelastic vibration mechanism in a thin ferromagnetic plate, when exposed to an external transverse magnetic field. The paper features a detailed analysis of…
Рақамли технологияларнинг назарий ва амалий масалалари — barcha maqolalar