Robust Audio Visual Biometric Person Authentication with Liveness Verification

In this paper we propose liveness verification for enhancing the robustness of audio-visual biometric person authentication systems. Liveness verification ensures that biometric cues are acquired from a live person who is actually present at the time of capture for authenticating the identity. The proposed liveness checking technique based on cross-modal association models involves hybrid fusion of acoustic and visual speech correlation features, which measure the degree of synchrony between the lips and the voice extracted from speaking face video sequences. Performance evaluation in terms of DET (Detector Error Tradeoff) curves and EERs (Equal Error Rates) on publicly available audiovisual speech databases show a significant improvement in robustness of system against different types of simulated replay attacks.