Robustness of Serial and Parallel Biometric Fusion against Spoof Attacks

Abstract

In this paper, we empirically study the robustness of multimodal biometric systems against spoof attacks. A few recent studies have questioned, contrary to a common claim in the literature, that a multimodal biometric system in parallel fusion mode can be cracked even if a single biometric trait is spoofed. Robustness of multimodal biometric systems in serial fusion mode against spoof attacks has so far not been investigated. We compare the performance of the multimodal systems with each mode under different spoof attack scenarios. We empirically show that multimodal biometric systems in both fusion modes are not intrinsically robust against spoof attacks as believed so far. In particular, multimodal biometric systems in serial fusion mode can be even less robust than systems in parallel mode, when only the best individual matcher is spoofed. Nonetheless, systems in serial fusion mode can be more robust than systems in parallel mode, when all matchers are spoofed.