How discernible is user impromptu behavior when unlocking a touch screen?

Being a popular computing device, smartphone (or tablet, wearable device) system's screen unlocking mechanisms have recently come under numerous successful attacks. In an aim to deter the attacks, we explored users' spontaneous touch behavioral traits in order to identify the users which we opine could be an added layer of security. We experimented with four conventional unlocking schemes: 4 Digit Pin, 3×3 “Connect the Dots”, Glow Pad, Slide Unlock and proposed one new unlocking mechanism, “Pressure Circle”. Six high-level features have been extracted from user touch activities and we designed our own verifier to identify users. For each unlocking scheme, experimental results generated from 992 impostor attacks performed on 32 user samples, show that users can be identified with an accuracy of up to 77.3% solely by considering their touch behavior. Overall, we achieve an average accuracy of 68.8%. We argue that, this inherent behavioral biometric trait when combined with the conventional unlocking schemes, would be less susceptible to attacks.