Stick, Swipe, Secure: Acoustic Cross-Device User Authentication via a Low-Cost Passive Fingerprint Sticker

With the widespread adoption of smart devices, traditional authentication methods are increasingly vulnerable to security risks. Recently, acoustic fingerprint authentication based on finger friction sound has gained attention as a cost-effective solution with distinctive acoustic properties. However, challenges persist in ensuring the distinctiveness, stability, and adaptability of acoustic fingerprints across devices. In this paper, we propose S 3 Guard , a user authentication system based on acoustic fingerprints. It uses a low-cost passive fingerprint sticker to physically standardize the acoustic environment during interaction. This approach amplify finger friction sound and combines the physiological features with user behavior characteristics to form a unique acoustic fingerprint for user authentication. The fingerprint sticker features a simple design, integrates easily into smart devices with microphones, and is nearly cost-free. We propose a FrictionES algorithm that efficiently separates clean finger friction sound from noise. Then, we employ a user authentication network model to dynamically extract features and perform authentication, integrating one-shot learning to tackle the cross-device transfer challenge. Compared to existing methods, S 3 Guard offers a lower-cost solution, stronger cross-device migration, effective amplification of acoustic features, and greater feature stability. Experimental results with 100 volunteers show that S 3 Guard achieves an average accuracy of 96.26%, effectively defending against spoofing, replay, and puppet attacks.