In the digital era, biometric authentication plays a crucial role in ensuring secure and seamless access to electronic services, particularly in payment systems where transaction security is paramount.Traditional authentication methods, such as passwords and PINs, have become increasingly vulnerable to cyber threats, necessitating the adoption of advanced biometric solutions.This study presents an improved authentication mechanism based on neural networks, specifically designed for face recognition in payment systems.The proposed approach integrates multiple deep learning models, including ResNet and Feature Pyramid Networks (FPN), to enhance accuracy and robustness.A novel two-factor authentication mechanism is introduced, incorporating emotion recognition to increase security against spoofing attacks.The system utilizes a lightweight neural network (LNN) optimized for real-time authentication, ensuring efficiency in practical implementations.Experimental results demonstrate a 95.7% face recognition accuracy and 88.9% emotion recognition accuracy, with an improved processing speed of 30 frames per second (FPS).The implementation of this mechanism in real-world payment systems shows significant potential for improving transaction security while maintaining user convenience.
| Jurnal | ТАТУ хабарлари |
|---|---|
| Nashr sanasi | 2025-08-19 |
| DOI | 10.61663/252tuitmct11 |
DOI: 10.61663/252tuitmct11 · Maqolaning asl sahifasi
Models in the form of ordinary differential equations traditionally have a wide range of applications.These models are widely applicable, including various mechanical systems, control systems, moving objects, and so…
The distribution of the temperature field of elements of threedimensional structures is examined in the article and the influence of nonhomogeneities on the temperature field in three-dimensional axisymmetric structures…
Algorithm and program for solving the gas filtration problem in dynamically interconnected porous media
The issue of managing reserves in the warehouse is considered urgent, and the provision of timely and sufficient production enterprise departments with them ensures the absence of production interruptions.The fact that…
Lemmatization is a vital task in Natural Language Processing (NLP), essential for various applications.However, the Uzbek language, categorized as a lowresource language in NLP, lacks dedicated lemmatization…
In today's world, the identification of first-degree kinship is considered crucial in issues such as family, inheritance, marriage, and confirming paternity or maternity.Despite the fact that research has been conducted…
Human action recognition is a fundamental task in the field of computer vision and has become increasingly important in applications such as human-computer interaction, intelligent surveillance systems, virtual and…
This article develops an IDEF0 model of the process of receiving and publishing an article by a user (author) on the uzjurnal.uzplatform, an IDEF0 model of the process of registering on the platform and submitting an…
This article develops a mathematical model and algorithm for determining and evaluating the similarity of web document objects.The proposed approach allows for assessing the functionality of web page objects, their…
As time progresses, the demand for digital data is increasing.As a graphic engineer, our task is to make a proposal that meets these requirements.We can provide information and data in various forms.The most optimal…