Gesture-related Two-factor Authentication for Wearable Devices via PPG Sensors

Verifying the user identity of wearable devices is crucial for system security, especially for sensitive operations like making financial payments. A PPG-based two-factor authentication can be a promising solution with widely deployed PPG (Photoplethysmography) sensors within wearable devices. Our observations find PPG readings reveal a significant relevance to the user's hand motions, i.e., gestures, while the user's heartbeat characteristics and wearing habits are also implicitly related, which can be utilized for user authentication. In this paper, we design G-PPG, a gesture-related PPG-based two-factor authentication mechanism that can non-intrusively validate the user's identity. In G-PPG, a gesture detection and segmentation module and a specific feature set are designed for accurate gesture-related PPG characteristic extraction. Moreover, a user-defined security level and an adaptive update scheme are proposed for the high accuracy of long-term authentication. The experiment results among 15 participants demonstrate that G-PPG can achieve an over 90% accuracy in different scenarios on average. With a carefully designed authentication mechanism, the pass rate for legitimate users can reach up to 96.67%, while the value is only 0.62% for attackers.