Continuous mobile authentication using a novel Graphic Touch Gesture Feature

TLDR

Behavioral biometrics on mobile devices are gaining attention, with touch gestures emerging as a promising biometric modality. The paper proposes a novel Graphic Touch Gesture Feature (GTGF) to extract identity traits from touch traces. GTGF encodes movement and pressure dynamics as intensity values and shapes, and its usability was evaluated on datasets of six common touch gestures. Combining six gestures yielded an equal error rate of 2.62%, demonstrating the method’s effectiveness.

Abstract

Behavioral biometric on mobile devices has begun to gain attention in recent years and the feasibility of touch gestures as a novel biometric modality has been investigated lately. In this paper, we propose a novel Graphic Touch Gesture Feature (GTGF) to extract the identity traits from the touch traces. The traces' movement and pressure dynamics are represented by intensity values and shapes of the GTGF. To evaluate its usability on the authentication problem, touch gesture datasets have been collected which includes six commonly used touch gestures. A Equal Error Rate of 2.62% has been achieved combining six gestures together, which demonstrated the effectiveness of the proposed methods.