BIOMETRIC AUTHENTICATION USING GENERATIVE AI TECHNIQUES

Abstract

Biometric authentication systems have become essential in modern security applications due to their ability to provide reliable and user-specific identification using physiological and behavioral traits such as face, fingerprint, iris, palmprint, voice, and vein patterns. Recently, generative artificial intelligence (AI) techniques, particularly Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), diffusion models, and transformer-based architectures, have significantly transformed biometric authentication by improving biometric data synthesis, spoof detection, privacy preservation, and multimodal authentication. This survey presents a comprehensive review of generative AI techniques applied in biometric authentication systems, covering major research areas including biometric data augmentation, deepfake detection, privacy-preserving authentication, multimodal biometric fusion, liveness detection, and anti-spoofing mechanisms. The survey is from 2020-2026.The study analyzes recent advancements in GAN-based synthetic biometric generation, adversarial attack detection, explainable AI integration, blockchain-enabled privacy protection, and quantum-resistant biometric frameworks. Furthermore, a detailed classification framework and comparative literature analysis are presented to identify current methodologies, datasets, algorithms, strengths, and limitations of existing approaches. The survey also highlights critical research gaps related to scalability, cross-database generalization, adversarial robustness, computational complexity, explainability, and real-time deployment challenges. Finally, future research directions are discussed, emphasizing the development of secure, lightweight, privacy-aware, and robust generative AI-driven biometric authentication systems for next-generation intelligent security applications.