A paper authored by Xinyue Niu (doctoral student) as the first author has been accepted for publication in IEEE Transactions on Biometrics, Behavior, and Identity Science.
Xinyue Niu and Akira Furui
Calibration-free Cross-Gesture EMG Biometrics for Personal Identification and Verification
IEEE Transactions on Biometrics, Behavior, and Identity Science (accepted).
This study proposes calibration-free cross-gesture EMG biometric authentication, in which enrollment and authentication gestures can be entirely different. The proposed DADA-Net disentangles identity and gesture representations through dual adversarial learning and conditional distribution alignment, yielding identity representations that generalize to unseen gestures. Experiments on two public datasets showed consistent gains over baselines, especially for unseen gestures, and revealed that wrist sEMG is well suited to authentication in consumer wearables.
This is Niu's first journal paper since joining our lab, and it will form part of Niu's doctoral dissertation. Congratulations!