Individual identification based on chaotic electrocardiogram signals

Electrocardiography (ECG) is a transthoracic interpretation of the electrical activity of the human's heart over time captured and is highly irregular, random, and variable from person to person. Recently, the literature has revealed that this kind of signal is, in fact, chaotic. Because of people's ECGs are extremely hard to be artificially duplicated, this paper intends to investigate the way of extracting the ECG signals' biometric features for the possibility of biometric recognition. The ECG signal is converted into the phase plane by using phase space reconstruction. Then, chaos extractor is applied to capture the representative indices of chaotic ECG signals i.e., correlation dimension and Lyapunov exponents spectrum. The root mean square, Lyapunov exponent and correlation dimension are used as the key variables in neural network training and utilized in the identification scheme.