In this paper we show the potential of two new features as powerful anti-spoofing measures for face-voice person authentication systems. The features based on latent semantic analysis (LSA) and canonical correlation analysis (CCA), enhance the performance of the authentication system in terms of better anti-imposture abilities and guard against video replay attacks, which is a challenging type of spoof attack. Experiments conducted on 2 speaking-face databases, VidTIMIT and UCBN, show around 42% improvement in error rate with CCA features and 61% improvement with LSA features over feature-level fusion of face-voice feature vectors.