*=Equal taxpayers
Personal devices have adopted various authentication methods, including biometric recognition and passwords. In contrast, headsets have limited input mechanisms, relying solely on the authentication of connected devices. We present Moonwalk, a novel method for passive user recognition using the built-in headphone accelerometer. Our focus is on gait recognition; allowing users to establish their identity by simply walking for a short interval, despite the sensor's location away from the feet. We employ self-supervised metric learning to train a model that produces a highly discriminative representation of a user's 3D acceleration, without the need for retraining. We tested our method in a study involving 50 participants, achieving an average F1 score of 92.9% and an equal error rate of 2.3%. We expanded our evaluation by evaluating performance in various conditions (e.g., shoe types and surfaces). We discuss the opportunities and challenges these variations introduce and propose new directions for advancing passive authentication for wearable devices.
