System and Method for Biometric Identification Using ECG Signals

Abstract Identification through a biometric can be built on behaviour or physiological traits of an individual. Since ECG captures an inherent physiological trait which can be used as an identity marker, the possibility of falsification and impostors can be reduced to a minimum. We investigate biometric identification systems based on ECG signals and their intra-subject and intra-session validity. In this invention, we develop a novel algorithm using Fourier decomposition method (FDM) and phase transform. The ECG signal is divided into frames consisting of one (or more) beats as input samples for the proposed system. These frames contain inter-beat information and also capture the intra-beat variations. They are decomposed into a set of Fourier intrinsic band functions using FDM and relevant features are extracted from them. Phase transform has been used here to highlight the intrinsic features hidden in the phase of ECG signals. The effects of variations in size of the frame, the decomposition levels, and the number of sessions used for training and testing, on the performance of the algorithm are analysed. Machine learning classifiers are used for identification. Figure 1: ECG signal acquisi on Remove R-peFrame detection and oudier fornton rce ction Phase Fourier I trasform decomposition transformmy n method Feature extraction Identification