Photoplethysmographic Subject Identification by Considering Feature Values Derived from Heartbeat and Respiration.

Area of Science:

• Biometrics
• Signal Processing
• Machine Learning

Background:

• Continuous authentication is crucial for security.
• Photoplethysmogram (PPG) signals offer a non-invasive data source.
• Existing biometric methods have limitations.

Purpose of the Study:

• To propose and evaluate a subject identification method using PPG signals.
• To leverage heartbeat and respiration features for authentication.
• To assess the method's accuracy on an open dataset.

Main Methods:

• Feature extraction from PPG signals, including heartbeat and respiration.
• Application of frequency filtering and Mel-Frequency Cepstrum Coefficients (MFCC).
• Generation of classifiers for performance evaluation.

Main Results:

• The proposed method achieved 92.9% accuracy in identifying 46 subjects.
• Feature values derived from heartbeat and respiration proved effective.
• Successful demonstration of PPG-based continuous authentication.

Conclusions:

• PPG signals can be effectively utilized for subject identification.
• The developed method offers a promising approach for continuous authentication.
• Heartbeat and respiration features are valuable for biometric recognition.