A HYBRID BIOMETRIC-CRYPTOGRAPHIC FRAMEWORK FOR SECURE ATM AUTHENTICATION USING FINGERPRINT RECOGNITION AND TIME-BOUND QR CODES

Abstract

The persistent vulnerabilities in traditional PIN-based ATM authentication systems, including skimming and shoulder surfing attacks, necessitate more robust security solutions. This paper presents a hybrid biometric-cryptographic framework for secure ATM authentication, combining fingerprint recognition with time-bound QR codes. The proposed framework addresses vulnerabilities in traditional PIN-based methods by implementing a dual-factor approach, combining fingerprint verification as the primary method with dynamically generated QR codes as a secure fallback. When fingerprint matching fails, the system generates a cryptographically signed QR code valid for 120 seconds, incorporating HMAC-SHA256 signatures and hardware-protected keys to prevent replay attacks. Experimental results using the Raspberry Pi 4 Model B demonstrate significant improvements, including 25% higher authentication accuracy than PIN systems and consistent processing times of 1.8-2.2ms for QR generation under load. The framework maintains usability while providing robust protection against skimming, shoulder surfing, and credential reuse. Key contributions include the integration of Hardware Security Module (HSM) protection for biometric templates, O(1) complexity QR validation, and automatic failover between authentication methods. This research offers financial institutions a practical and scalable approach to enhance ATM security, eliminating the need for significant infrastructure modifications.