Touchless Biometrics – Capture and Matching (TB-CAM)

The TB-CAM Phase II effort aims to develop a mobile-based multifactor authentication system that integrates facial, periocular, and touchless fingerprint biometrics for secure and efficient authentication in military environments. Designed for deployment on Android-based Army devices, TB-CAM leverages machine learning-driven biometric recognition models optimized for real-time performance in denied, degraded, intermittent, and limited (DDIL) environments. The system dynamically adapts authentication methods based on user occlusions, environmental conditions, and operational constraints, ensuring reliable identity verification even when soldiers wear protective gear. Occluded facial recognition capabilities, developed in collaboration with Clarkson University under the Face1st project, will enable authentication even when full facial visibility is obstructed. Additionally, TB-CAM will incorporate touchless fingerprint recognition, utilizing deep learning models to enhance non-contact-to-contact fingerprint matching and maintain compatibility with legacy biometric databases. Advanced Presentation Attack Detection (PAD) methods will safeguard against spoofing, including deepfake, 3D mask, and synthetic fingerprint attacks. By leveraging secure enclave processing and AES-256 encryption, TB-CAM ensures robust data protection. This effort will result in a field-ready prototype, demonstrating compatibility with Nett Warrior and ATAK platforms and setting the stage for Phase III transition and broader defense and commercial applications.