Robustness Testing of Artificial Intelligence-Driven Hand Biometric-Based User Authentication in Mobile Devices

User authentication has become increasingly important to secure data digitally. The proposed solution emits an acoustic signal from a device to scan individual hand biometrics for access. This involved sound propagation through a user's hand and analyzing the effects that background noise, hand moisture, hand motion, and hand position have on propagation. MATLAB was used to train and test artificial intelligence models to test robustness. This study found that recordings affected by any of the variables had average validation accuracies of about 90 %. The average validation and test accuracies between different users' baselines were 88 % and 52 %, respectively.