Innovative Deep Learning Architecture for Enhanced Altered Fingerprint Recognition

2025-09-24

AI Summary

To address fingerprint recognition failure caused by deliberate fingerprint alterations—such as those employed to evade border control or forensic identification—this paper proposes DeepAFRNet, the first model rigorously evaluated on the real-world tampered fingerprint dataset SOCOFing under three difficulty levels (Easy/Medium/Hard). The model adopts VGG16 as its backbone for discriminative feature extraction and integrates cosine similarity–based embedding matching with stringent thresholding to enhance robustness. Unlike prior approaches reliant on synthetic data, DeepAFRNet operates directly on authentic altered fingerprints. Experiments demonstrate classification accuracies of 96.7%, 98.76%, and 99.54% across the three difficulty tiers, respectively—marking substantial improvements in recognition reliability under adversarial manipulation. These results validate DeepAFRNet’s high robustness and practical deployability for security-critical applications.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Intelligent Robots: Manipulation

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Altered fingerprint recognition (AFR) is challenging for biometric verification in applications such as border control, forensics, and fiscal admission. Adversaries can deliberately modify ridge patterns to evade detection, so robust recognition of altered prints is essential. We present DeepAFRNet, a deep learning recognition model that matches and recognizes distorted fingerprint samples. The approach uses a VGG16 backbone to extract high-dimensional features and cosine similarity to compare embeddings. We evaluate on the SOCOFing Real-Altered subset with three difficulty levels (Easy, Medium, Hard). With strict thresholds, DeepAFRNet achieves accuracies of 96.7 percent, 98.76 percent, and 99.54 percent for the three levels. A threshold-sensitivity study shows that relaxing the threshold from 0.92 to 0.72 sharply degrades accuracy to 7.86 percent, 27.05 percent, and 29.51 percent, underscoring the importance of threshold selection in biometric systems. By using real altered samples and reporting per-level metrics, DeepAFRNet addresses limitations of prior work based on synthetic alterations or limited verification protocols, and indicates readiness for real-world deployments where both security and recognition resilience are critical.

Problem

• Recognizing deliberately altered fingerprints to prevent evasion of biometric systems
• Matching distorted fingerprint samples using deep learning feature extraction
• Addressing limitations of synthetic alterations with real altered fingerprint evaluation

Innovation

• Uses VGG16 backbone for high-dimensional feature extraction
• Employs cosine similarity for fingerprint embedding comparison
• Achieves high accuracy on real altered fingerprint datasets