Biometric Authentication using ECG Signals

An electrocardiogram commonly called ECG or EKG is an electrical signal from the heart which is generally used to monitor the health of the heart. Due to its highly discriminatory property, ECG-based authentication has recently achieved great attention. However, there are still many challenges in discriminatory feature extraction and efficient pattern classification presented in the state of the arts and various research studies. We initially removed the noise from the ECG signals consisting of noise components of high and low frequencies with the help of the moving average filter. The process of extracting necessary parts of the input signals which contain discriminatory information related to each individual is performed by use of the Empirical Mode Decomposition method. Thereafter the feature extraction is carried out with the features of the statistical domain, time domain, and frequency domain taken from the decomposed signals of ECG data. The classification is done using the selected features from ECG signals data with various algorithms for biometric authentication to know the accuracy and performance of each method and for choosing the most effective classification.