ECG authentication in post-exercise situation.

Area of Science:

• Biometrics
• Cardiovascular Physiology
• Signal Processing

Background:

• Human authentication using physiological signals is a growing research area.
• Electrocardiogram (ECG) offers a unique and non-invasive biometric modality.
• The impact of physiological stress on ECG-based authentication requires further investigation.

Purpose of the Study:

• To evaluate the feasibility of human authentication using ECG signals acquired after strenuous physical activity.
• To determine the optimal duration of post-exercise ECG recording for reliable subject recognition.

Main Methods:

• ECG signals were collected from 55 subjects before and after a 5-minute stepper exercise.
• Linear Discriminant Analysis (LDA) was employed for feature extraction and classification.
• Analysis focused on ECG data segments of varying durations post-exercise.

Main Results:

• Subject recognition accuracy was low (<60%) within the first minute post-exercise.
• Accuracy rapidly increased to over 90% after one minute of rest.
• Peak accuracy reached 96.22% within 5 minutes post-exercise.

Conclusions:

• ECG-based human authentication is effective when utilizing data acquired after a brief recovery period post-exercise.
• A minimum of one minute of post-exercise ECG recording is recommended for reliable authentication.
• This method presents a promising approach for biometric security in real-world applications.