Investigating Mobile Device Picking-up motion as a novel biometric modality

Employing mobile sensor data to recognize user behavioral activities has been well studied in recent years. However, exploiting mobile motion data as a novel biometric modality remains a new area. In this paper, we propose two novel methods, a Statistic Method to intuitively apply classifier on the statistic features of the data; and a Trajectory Reconstruction Method to reconstruct the Mobile Device Picking-up(MDP) motion trajectories and extract specific identity features from the traces. We evaluated our methods on a multi-session motion dataset. A Equal Error Rate of 6.13% and 7.09% has been respectively achieved by the Statistic Method and the Trajectory Reconstruction Method, which demonstrated the feasibility of the proposed methods. Furthermore, experimental results showed several interesting evidences: 1) the accuracy of the methods declined in the inter-session tests; and 2) user movements(e.g., walking) have a high impact on the verification performance.