Abstract:<p>Traditional fingerprint recognition systems are vulnerable to attacks, such as the use of artificial fingerprints, and poor performance will be achieved if the captured surface fingerprints are of low-quality. Developing high-security and robust fingerprint recognition systems is of increasing concern in modern society. The introduction of optical coherence tomography (OCT) for fingerprint imaging opens up a new research domain for fingerprint recognition due to its ability to capture the depth information of skin layers. This paper proposes a fingerprint recognition system based on OCT. The research first establishes a database with normal, worn-out, artificial and degraded fingerprints imaged by our custom-built, in-house OCT device. Then, we propose to reconstruct three layers of subsurface fingerprints by considering diverse skin layer information. For each of the subsurface fingerprints divided according to the physical structure of the fingertip, we propose a simple yet effective projection-based reconstruction method. Finally, a pixel-level fusion strategy based on the local image quality is proposed to the fuse the three levels of subsurface fingerprints for robust fingerprint recognition. In our experiments, we show that as much diverse fingerprint information can be retained as possible by the proposed subsurface fingerprint reconstruction method. We also demonstrate the effectiveness and efficiency of the proposed method by comparing it with existing state-of-the-art internal fingerprint reconstruction approaches. The robustness and high anti-spoofing ability of the proposed system is verified by comparing the matching performance evaluated on the established OCT-based database and another database with the same fingerprints imaged by a commercial optical sensor. The best EER and FMR100 evaluated on OCT-based fingerprints are 0.42% and 0.36%, respectively. The best EER and FMR100 evaluated on traditional 2D surface fingerprints are 8.05% and 18.18%, respectively, which shows the vast potential of the proposed system in current automated fingerprint recognition systems (AFRSs).</p>

A New Approach in Automated Fingerprint Presentation Attack Detection Using Optical Coherence Tomography

Internal Structure Attention Network for Fingerprint Presentation Attack Detection from Optical Coherence Tomography

A Uniform Representation Learning Method for OCT-based Fingerprint Presentation Attack Detection and Reconstruction

Prototype-Guided Autoencoder for OCT-Based Fingerprint Presentation Attack Detection

OCT Fingerprints: Resilience to Presentation Attacks

Fingerprint Presentation Attack Detection by Region Decomposition

One-Class Fingerprint Presentation Attack Detection Using Auto-Encoder Network

Fingerprint Presentation Attack Detection by Channel-Wise Feature Denoising

Presentation Attack Detection with Advanced CNN Models for Noncontact-based Fingerprint Systems

Reliable internal fingerprint detection using micro-optical coherence tomography

A uniform representation model for OCT-based fingerprint presentation attack detection and reconstruction

Biometric Presentation Attack Detection: Beyond the Visible Spectrum

Fingerprint Presentation Attack Detector Using Global-Local Model

Fingerprint Presentation Attack Detection by Learning in Frequency Domain

Fingerprint Presentation Attack Detection: A Sensor and Material Agnostic Approach

Robust and high-security fingerprint recognition system using optical coherence tomography

A Survey on Unknown Presentation Attack Detection for Fingerprint

Fingerprint Presentation Attack Detection Based on Local Features Encoding for Unknown Attacks

Contrast Enhanced Subsurface Fingerprint Detection Using High-Speed Optical Coherence Tomography

On the Generalisation Capabilities of Fingerprint Presentation Attack Detection Methods in the Short Wave Infrared Domain

Taming Self-Supervised Learning for Presentation Attack Detection: De-Folding and De-Mixing