Biometric Authentication System Based on Electrocardiogram (ECG)

Electrocardiogram (ECG) is the measure of the electrical activity of heart which provides a powerful authentication system for person recognition due to the variation in ECG signatures. In spite of fingerprint biometrics and face detection in past years, biometric system based on ECG is more in pace. In this article a simple method is proposed for person recognition consisting of first denoising the signal by using wavelet decomposition. The denoised signal is then passed through the process of EMD. First two IMFs are selected and from the resulting signal a combination of time domain and frequency domain features are extracted and cubic SVM is trained to get a highest accuracy of 98.4\% with 35 subjects.