Dynamic Liveness-Integrated CNN Architecture for Face-Iris Spoof Detection

• Author(s): Afolabi Awodeyi ; Philip Asuquo ; Bliss Stephen
• Paper ID: 1710350
• Page: 1079-1089
• Published Date: 01-09-2025
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 9 Issue 2 August-2025

Abstract

This study addresses the critical vulnerability of biometric systems to spoofing and presentation attacks by proposing a secure, CNN-integrated anti-spoofing framework for multimodal face and iris recognition. The purpose of the research is to enhance the reliability of biometric authentication systems by embedding dynamic liveness detection directly into the recognition pipeline. The system leverages physiological cues such as eye blinking, temporal texture variations, and pupil motion, which are difficult to replicate in spoof media like printed images and video replays. A dual-branch CNN architecture is employed to extract both spatial and temporal features from face and iris inputs. These features are fused with liveness indicators to improve decision accuracy. Synthetic spoofing attacks were introduced into the ORL and CASIA-IrisV4 datasets to simulate real-world adversarial conditions, including high-resolution photo and video-based attacks. The model was trained and evaluated using standard biometric security metrics. Results show that the proposed system achieves a spoof detection accuracy of 98.9%, with a false acceptance rate (FAR) of 0.00% and a false rejection rate (FRR) of 1.1%. Compared to baseline CNN models without liveness integration, the framework significantly improves resilience against spoofing. In conclusion, the integration of dynamic liveness cues into CNN-based multimodal recognition enhances both security and reliability, making the system suitable for deployment in high-assurance access control environments and offering a foundation for future mobile or edge-based biometric solutions.

Keywords

Anti-spoofing, Liveness detection, CNN, Face-Iris biometrics, Presentation attack detection

Citations

IRE Journals:
Afolabi Awodeyi , Philip Asuquo , Bliss Stephen "Dynamic Liveness-Integrated CNN Architecture for Face-Iris Spoof Detection" Iconic Research And Engineering Journals Volume 9 Issue 2 2025 Page 1079-1089

IEEE:
Afolabi Awodeyi , Philip Asuquo , Bliss Stephen "Dynamic Liveness-Integrated CNN Architecture for Face-Iris Spoof Detection" Iconic Research And Engineering Journals, 9(2)