A novel biometric authentication approach using ECG and EMG signals

Area of Science:

• Biometrics and Physiological Signal Processing
• Cybersecurity and Identity Verification
• Medical Signal Analysis

Background:

• Traditional identity verification methods like passwords are vulnerable.
• Electrocardiogram (ECG) signals, characterized by P, Q, R, S, and T waves, are unique to individuals.
• ECG biometrics offers inherent liveness detection, a capability lacking in other biometric systems.

Purpose of the Study:

• To present a low-cost, online system for acquiring and processing ECG signals for personal authentication.
• To explore the fusion of ECG and Electromyogram (EMG) signals for enhanced biometric system reliability.
• To investigate the potential for retrieving personal data from ECG signals for improved decision-making.

Main Methods:

• A novel biometric system combining ECG and EMG physiological recordings was developed.
• ECG signals were collected using a minimally invasive, palm-based, one-lead setup.
• Feature extraction utilized Fourier descriptors, and the Optimum-Path Forest classifier was employed for similarity calculations.

Main Results:

• The study exploited five public ECG databases and in-house collected signals, analyzing 3750 ECG beats.
• The fused ECG/EMG biometric system demonstrated robust subject authentication capabilities.
• The proposed approach yielded promising results for reliable individual authentication.

Conclusions:

• ECG and EMG signal fusion presents a novel and effective biometric authentication method.
• The developed palm-based ECG system offers a low-cost solution for identity verification.
• The findings suggest significant potential for robust and secure personal authentication using physiological signals.