Unobservable Re-authentication for Smartphones.

TLDR

Smartphones are widely used personal devices that store private data and serve as network entry points, making them attractive targets for theft and impersonation attacks that threaten user security and privacy. This work proposes a continuous, unobtrusive biometric re‑authentication system for smartphones to verify the current user’s identity and notify the owner when needed. The system trains a classifier on the owner’s finger‑movement patterns and continuously compares incoming patterns to the model, enabling real‑time, non‑intrusive re‑authentication without disrupting user interactions. Experimental results demonstrate that the system runs efficiently on smartphones and achieves high authentication accuracy.

Abstract

The widespread usage of smartphones gives rise to new security and privacy concerns. Smartphones are becoming a personal entrance to networks, and may store private information. Due to its small size, a smartphone could be easily taken away and used by an attacker. Using a victim’s smartphone, the attacker can launch an impersonation attack, which threatens the security of current networks, especially online social networks. Therefore, it is necessary to design a mechanism for smartphones to re-authenticate the current user’s identity and alert the owner when necessary. Such a mechanism can help to inhibit smartphone theft and safeguard the information stored in smartphones. In this paper, we propose a novel biometric-based system to achieve continuous and unobservable re-authentication for smartphones. The system uses a classifier to learn the owner’s finger movement patterns and checks the current user’s finger movement patterns against the owner’s. The system continuously re-authenticates the current user without interrupting user-smartphone interactions. Experiments show that our system is efficient on smartphones and achieves high accuracy.