A Novel Person Authentication Technique Using Electrocardiogram (ECG)

Abstract

Since the beginning of this century, electrocardiogram (ECG) signals have garnered a growing amount of attention for the purpose of person identification due to the distinctive qualities and physiological significance they possess. Nevertheless, electrocardiogram (ECG) data are often tainted by noise, which may lead to a decrease in the efficiency of authentication systems. In order to successfully denoise the input electrocardiogram (ECG) data, the CEEMDAN-NLMS filter is used. This filter eliminates undesired noise while maintaining the essential characteristics.  The proposed CNN architecture is responsible for providing the spatial and temporal dependencies that are present within the ECG data. Our approached technique, deep learning CNN model with machine learning SVM algorithm delivers greater performance in terms of authentication accuracy, which is around 99.25%, surpassing methods that are considered to be state-of-the-art and demonstrating that our method is superior. Overall, our research demonstrates an innovative and efficient method for enhancing the authenticity of individuals via the use of electrocardiogram (ECG).