A Biometric Identification System Based on Eigenpalm and Eigenfinger Features
IEEE Transactions on Pattern Analysis and Machine Intelligence
Verification of humans using the electrocardiogram
Pattern Recognition Letters
eigenPulse: Robust human identification from cardiovascular function
Pattern Recognition
Pattern Recognition
Evaluating the use of ECG signal in low frequencies as a biometry
Expert Systems with Applications: An International Journal
Multimodal biometric system combining ECG and sound signals
Pattern Recognition Letters
Hi-index | 0.10 |
Most electrocardiogram (ECG) biometrics are based on detection of the QRS wave and comparison of structural features. ECG parameters are extracted from the waveform, but the process is arduous for noisy signals. Comparison based on phase space trajectory from a cardiac cycle as well as waveform comparison avoids the detection of ECG characteristic points, but has an alignment-free advantage. In this paper, we develop a QRS detection-free ECG biometric based on the phase space trajectory of the ECG signal. The multi-loop trajectory from a 5-s ECG epoch is condensed to a single-loop coarse-grained structure. The normalized spatial correlation (nSC), the mutual nearest point match (MNPM), and the mutual nearest point distance (MNPD) are considered as means of quantifying the similarity or dissimilarity between coarse-grained structures. We test our method on a population of 100 subjects. The accuracies of personal identification achieved for a single-lead ECG are 96%, 95%, and 96% for the MNPD, nSC, and MNPM methods respectively. When we analyze the phase space trajectory of a three-lead ECG, the accuracies increase to 99%, 98%, and 98% respectively. The coarse-grained phase space trajectory of an ECG signal is unambiguous and easy to compute, rendering ECGs a practical alternative to other biometrics.