Bi-Encoder Contrastive Learning for Fingerprint and Iris Biometrics

2025-10-26

AI Summary

This study challenges the conventional assumption of statistical independence among biometric traits by systematically investigating cross-modal correlations between fingerprints and irises. Method: We propose a dual-encoder contrastive learning framework built upon ResNet-50 and Vision Transformer backbones, enabling end-to-end modeling of intrinsic inter-modal associations across multi-biometric traits from the same individual. Contribution/Results: Experiments reveal statistically significant correlation between left and right irises (p < 0.01), and—critically—cross-modal fingerprint–iris matching performance, though limited, significantly exceeds chance level (p < 0.01), thereby undermining the foundational premise of biometric trait independence. Iris intra-modal matching achieves 91% ROC AUC; fingerprint models replicate strong intra-individual consistency. These findings provide novel empirical evidence and a methodological paradigm for multi-modal biometric recognition, privacy risk assessment, and generative biometric modeling.

Technology Category

• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning

Application Category

• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

There has been a historic assumption that the biometrics of an individual are statistically uncorrelated. We test this assumption by training Bi-Encoder networks on three verification tasks, including fingerprint-to-fingerprint matching, iris-to-iris matching, and cross-modal fingerprint-to-iris matching using 274 subjects with $sim$100k fingerprints and 7k iris images. We trained ResNet-50 and Vision Transformer backbones in Bi-Encoder architectures such that the contrastive loss between images sampled from the same individual is minimized. The iris ResNet architecture reaches 91 ROC AUC score for iris-to-iris matching, providing clear evidence that the left and right irises of an individual are correlated. Fingerprint models reproduce the positive intra-subject suggested by prior work in this space. This is the first work attempting to use Vision Transformers for this matching. Cross-modal matching rises only slightly above chance, which suggests that more data and a more sophisticated pipeline is needed to obtain compelling results. These findings continue challenge independence assumptions of biometrics and we plan to extend this work to other biometrics in the future. Code available: https://github.com/MatthewSo/bio_fingerprints_iris.

Problem

• Testing statistical correlation between different biometric modalities
• Developing cross-modal fingerprint-to-iris matching using Bi-Encoder networks
• Challenging independence assumptions in biometric systems

Innovation

• Bi-Encoder networks for cross-modal biometric matching
• Contrastive learning minimizes same-subject image differences
• Vision Transformers applied to biometric matching tasks