Person identification in irregular cardiac conditions using electrocardiogram signals

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Electrocardiogram (ECG) signals possess unique characteristics that can be leveraged for biometric identification.
• Irregular cardiac conditions present unique challenges and opportunities for ECG-based biometrics.

Purpose of the Study:

• To develop and evaluate a person identification mechanism using ECG signals in individuals with irregular cardiac conditions.
• To assess the efficacy of cross-correlation as a biometric matching algorithm for ECG signals from abnormal cardiac databases.

Main Methods:

• Utilized ECG data from 30 subjects across three public databases: AFPDB, SVDB, and TWADB, encompassing various abnormal heart conditions.
• Employed cross-correlation (CC) as the primary biometric matching algorithm.
• Evaluated performance using False Acceptance Rate (FAR) and False Reject Rate (FRR) metrics.

Main Results:

• Achieved high recognition rates across all tested abnormal cardiac databases.
• Reported low False Acceptance Rates (FAR) of 2%, 3%, and 2% for AFPDB, SVDB, and TWADB, respectively.
• Reported low False Reject Rates (FRR) of 1%, 2%, and 0% for AFPDB, SVDB, and TWADB, respectively.

Conclusions:

• ECG-based biometric identification is effective for individuals with irregular cardiac conditions.
• The QRS complex within ECG morphology contains salient biometric characteristics suitable for individual differentiation.
• Cross-correlation proves to be a simple yet effective algorithm for ECG-based person identification in abnormal cardiac states.