Vitality Detection from Biometrics: State-of-the-Art Yogendra Narain Singh Department of Computer Science & Engineering Institute of Engineering & Technology, Gautam Buddh Technical University Lucknow, India Email: singhyn@gmail.com Sanjay Kumar Singh Department of Computer Engineering Institute of Technology, Banaras Hindu University Varanasi, India Email: sks@itbhu.ac.in Abstract—Biometric authentication systems have become the basis of trust to human society from the beginning of last decade, but the security of the system can be breached by presenting a non-live or fake biometric sample which is the cloned of a legitimate user’s identity. Different techniques have been proposed to address the problem of vitality detection from biometrics but these techniques are far away from the final solution. This paper proposes a classification of vitality detection techniques for fingerprint, face and iris biometrics for summarizing the current state-of-the-art and presents a critical review of them. We evaluate the potential of multimodal techniques of vitality detection from biometrics and analyze the performance of different vitality detection techniques on the datasets and the test conditions that have used. The use of physiological signals that have the inherited feature of vitality signs as a supplementary information with other conventional biometrics can offer a potential defense against spoofing attacks in the system. Keywords-Biometrics; Vitality Detection; Spoofing Attacks; Fake Biometrics; Security; I. I NTRODUCTION Biometrics refer a technology to authenticate individuals by automated means that rely on anatomical or behavioural human characteristics [1]. Biometric systems have the po- tential to do the people authentication with a high degree of assurance. Many body parts and personal characteristic have been suggested and used for biometrics systems e.g., face, fingerprint, handgeometry, iris, palmprint, voice and handwritten signatures. Biometrics are unique identifiers but they are not really the secrets e.g., fingerprints are left on everything to touch, facial geometry and iris patterns are visible while voices are being recorded. A sample of biometric can be acquired covertly by synthetic reproduction of anatomical identities e.g., acquisi- tion of facial and iris images, lifting of latent fingerprints, or imitation of behavioral identity e.g., producing similar voice and prepare the digital artifacts. The digitized artifact is the cloned of a legitimate user’s identity. The digital clone then be presented to the biometric system to get access as an legitimate individual and deceive the system. The most susceptible point in a biometric system of secu- rity vulnerability is the sensor interface where people present their biometrics to prove their identity. Sensors are the easy Feature Extractor Fake Biometric Sample Template Database Matcher No Yes Hardware Software Data Processing Replay Attacks/ False Data Inject Sensors Administrator Figure 1: Attacks of non-live or fake biometrics on a typical biometric system. target for intruders to present the fake biometric samples and spoof the system. A replay of the stored information or false data can also be injected in the place of processing chain to circumvent the system as shown in Figure 1. Therefore, the act of differentiating a genuine biometric sample acquired from a live person and the data acquired from other source is the aim of vitality detection. Fingerprint, face and iris are the common biometrics used by most of the present authentication systems. As per our knowledge, no previous work is reported in the literature to review state-of-the-art of vitality detection techniques from these biometrics, conjointly. In this paper we propose the classification of vitality detection techniques of fingerprint, face and iris biometrics for summarizing current state-of-the- art and present a critically review in Section II. The potential of multimodal techniques of vitality detection are evaluated in Section III. Performance of the cited vitality detection techniques are compared on the datasets and the material used to prepare fake biometrics in Section IV. An overview of physiological signals as emerging biometrics that have the inherited feature of vitality signs is given in Section V. Finally, the conclusion is drawn in Section VI. 106 978-1-4673-0126-8/11/$26.00 c 2011 IEEE