The Morphology of the Electrocardiogram for Eevaluating ECG Biometrics

Abstract

The morphology of the electrocardiogram (ECG) deals with interpretation of the ECG waveforms. The goal of this research is to evaluate whether ECG biometric system can be fooled by the morphology of the electrocardiogram. Two ECG morphology methods were investigated for their performance, which are McSharry's ECG dynamical model with genetic algorithm and P,Q,R,S,T fractal point interpolation method. Our experiment evaluated system performances by using the mean square error and the correlation coefficient, and our results showed PQRST fractal point interpolation morphology method may have better chance to fool current feature-based ECG biometric system if the locations and amplitudes of PQRST points has been known or guessed. Then, the correlation-based ECG biometric system has better chance to survive. However, it is still hard to mimic some details of ECG signals such as an angle of the S wave, and T wave downward. Hence, the future ECG biometric system is suggested by taking small details into account to prevent the possible false.