Hybrid Biometric Systems for Enhanced Security in IoT and Cloud-Based Applications Using Deep Learning and Signal Processing

Abstract

The fast growth of the Internet of Things (IoT) and cloud computing services has led to the need to have sophisticated security architectures to ensure that sensitive user information is secured across a network of distributed devices. The single-factor authentication method, e.g., passwords or basic tokens, is becoming more susceptible to advanced cyber-attacks, such as phishing, credential stuffing, and brute-force attacks. The hybrid biometric authentication system suggested in this study will combine physiological characteristics, deep learning, and signal processing methods to offer a high-security level. The study builds on a multimodal solution, which means integrating facial recognition and signal-based biometrics to address the shortcomings of unimodal solutions to vulnerabilities to spoofing and presentation attacks, as well as external noise. The hybrid deep learning architecture that is used in the methodology is that of the Convolutional Neural Networks (CNN), which is used to extract features and process the signal to detect liveness and reduce noise. The experiments on virtual data show that the given hybrid model is the most effective one, with an accuracy of authentication of 98.6, which is much higher than the results of the traditional methods, with an average of 90-92. In addition, the system has a 15.8 percent lower rate of false acceptance than the old structures. The statistical analysis proves that signal-to-noise ratios of biometric data are minimized by the integration of signal processing by 12 percent, which increases the reliability of the signal processing in edge clouds. This study finds that hybrid biometric systems optimized through deep learning offer a scalable, secure, and user-friendly solution to the continuously changing IoT. Based on the current peer-reviewed sources, this study creates a standard of multi-step authentication algorithms in securing data storage and real-time identity control in high-stakes applications such as healthcare and banking. The framework proposed shows resistance to adversarial threats of the present-day world.