Audio-Visual Biometrics and Forgery

Abstract

With the emergence of smart phones and third and fourth generation mobile and communication devices, and the appearance of a "first generation" type of mobile PC/PDA/phones with biometric identity verification, there has been recently a greater attention to secure communication and to guaranteeing the robustness of embedded multi-modal biometric systems. The robustness of such systems promises the viability of newer technologies that involve e-voice signatures, e-contracts that have legal values, and secure and trusted data transfer regardless of the underlying communication protocol. Realizing such technologies require reliable and error-free biometric identity verification (IV) systems. Biometric IV systems are starting to appear on the market in various commercial applications. However, these systems are still operating with a certain measurable error rate that prevents them from being used in a full automatic mode and still require human intervention and further authentication. This is primarily due to the variability of the biometric traits of humans over time because of growth, aging, injury, appearance, physical state, and so forth. Imposture can be a real challenge to biometric IV systems. It is reasonable to assume that an impostor has knowledge of the biometric authentication system techniques used on one hand, and, on the other hand, has enough information about the target client (face image, video sequence, voice recording, fingerprint pattern, etc.) A deliberate impostor attempting to be authenticated by an IV system could claim someone else’s identity to gain access to privileged resources. Taking advantage of the non-zero false acceptance rate of the IV system, an impostor could use sophisticated forgery techniques to imitate, as closely as possible, the biometric features of a genuine client. The robustness of a biometric IV system is best evaluated by monitoring its behavior under impostor attacks. This chapter studies the effects of deliberate forgery on verification systems. It focuses on the two biometric modalities people use most to recognize naturally each other: face and voice. The chapter is arranged as follows. Section 2 provides a motivation of the research investigated in this chapter, and justifies the need for audio-visual biometrics. Section 3 introduces audio-visual identity verification and imposture concepts. Section 4 then reviews automated techniques of audio-visual (A/V) IV and forgery. Typically, an A/V IV system uses audio and video signals of a client and matches the features of these signals with stored templates of features of that client. Next, section 5 describes imposture techniques on the visual and audio levels: face animation and voice conversion. The video sequence of the client can be altered at the audio and the 3