A novel biometric authentication approach using electrocardiogram signals

Area of Science:

• Biometrics
• Signal Processing
• Medical Informatics

Background:

• Biometric authentication is crucial for secure access.
• Electrocardiogram (ECG) signals offer unique physiological markers for identification.
• Existing ECG-based methods using AC/DCT or MFCC features have limitations.

Purpose of the Study:

• To propose a novel, robust biometric authentication system.
• To enhance recognition accuracy by combining multiple ECG signal features.
• To evaluate the proposed system's performance against existing methods.

Main Methods:

• Feature extraction using AC/DCT and MFCC from ECG signals.
• Inclusion of QRS beat information for enhanced discriminative power.
• System validation on a diverse subset of 30 subjects from the PTB database.

Main Results:

• The proposed combined feature approach achieved an average frame recognition rate of 97.31%.
• Performance evaluation demonstrated superiority over systems using AC/DCT or MFCC features alone.
• The system showed effective authentication capabilities on both healthy and non-healthy subjects.

Conclusions:

• Combining AC/DCT, MFCC, and QRS beat information significantly improves ECG-based biometric authentication.
• The novel approach offers a more accurate and reliable biometric solution.
• This method presents a promising advancement in physiological-based security systems.