Biometric authentication via keystroke sound

Unlike conventional “one shot” biometric authentication schemes, continuous authentication has a number of advantages, such as longer time for sensing, ability to rectify authentication decisions, and persistent verification of a user´s identity, which are critical in applications demanding enhanced security. However; traditional modalities such as face, fingerprint and keystroke dynamics, have various drawbacks in continuous authentication scenarios. In light of this, this paper proposes a novel nonintrusive and privacy-aware biometric modality that utilizes keystroke sound. Given the keystroke sound recorded by a low-cost microphone, our system extracts discriminative features and performs matching between a gallery and a probe sound stream. Motivated by the concept of digraphs used in modeling keystroke dynamics, we learn a virtual alphabet from keystroke sound segments, from which the digraph latency within pairs of virtual letters as well as other statistical features are used to generate match scores. The resultant multiple scores are indicative of the similarities between two sound streams, and are fused to make a final authentication decision. We collect a first-of-its-kind keystroke sound database of 45 subjects typing on a keyboard. Experiments on static text-based authentication, demonstrate the potential as well as limitations of this biometric modality.