ABSTRACT

This paper presents a method to analyze electrocardiogram (ECG) signal, extract the features, for the real time human identification. Data were obtained from short-term Lead-I ECG records (only one lead) of forty students at Paris Est University (UPEC). Signal averaging was applied to generate ECG databases and templates for reducing the noise recorded with palm ECG signals. Time domain signal processing is performed, which comprises the usual steps of filtering, peak detection, heartbeat waveform segmentation, and amplitude normalization. In this paper, an ECG biometric recognition method, that needs detection of one fiducial point only is introduced, based on classification of coefficients from the Fast Fourier Transform (FFT) of the Autocorrelation (AC) sequence of ECG data segments. The FFT is used to reduce extracted features from ECG signals. A 100% subject recognition rate for healthy subjects can be achieved for parameters that are suitable for the database.