Utter Innocence: Contactless Authentication Method Based on Physiological Signals

User authentication is crucial for privacy and security. In daily life, personalized intelligent devices are frequently used; however, these devices are vulnerable to unauthorized access, which can result in serious privacy breaches. To address this problem, this article proposes a user security authentication method that uses millimeter-wave radar to perceive human life signals for verification. First, the millimeter-wave radar is used to perceive the user’s breathing and heartbeat signals. Second, to eliminate the influence of breath-holding on the verification result, this article adopts the method of recognizing breathing patterns and establishing two channels: one for other breathing signals when not holding breath and another for heartbeat signals when holding breath. Finally, the wavelet packet decomposition (WPD) and Mel-frequency cepstral coefficients (MFCC) features of the user’s breathing signals, as well as the MFCC features of the heartbeat signals, are extracted. A lightweight neural network, ShuffleNet V2, is used for security authentication. Through a large number of experiments, the verification accuracy is 93.7%.