|
With the rapid proliferation of smartwatch, a secure and convenient smartwatch-based user authentication scheme are desired. As the widely deployed bioelectrical signal sensor in smartwatch, Photoplethysmography (PPG) sensors have shown potentials for authentication. Existing authentication solutions usually have some limitations. They require the user to provide an amount of registration data from user to reflect the profile of user, which may impact the experience of user. In this paper, we propose a PPG-based smartwatch authentication scheme. We leverage the Siamese Network to extract the feature of user from the PPG signal affected by the finger-level gesture for authentication. We conduct some experiments to evaluate the performance of the scheme. The experiment results show that our model has an average accuracy rate of 92.43\%. In addition, the authentication model can achieve high authentication accuracy with a small amount of user registration data. |
|
Keywords:User authentication, smartwatch, PPG sensor,siamese network |
|