ECG BASED BIOMETRIC FOR DOUBLY SECURE AUTHENTICATION

This paper presents a new ECG based biometric system that has a doubly secure characteristic when applied to authentication. In this system the parameters of the new underlying pulse active ratio (PAR) feature vector process are represented by a 4 digit security personal identification number (PIN). An authentication is achieved by firstly using the correct PIN and, secondly, by obtaining a positive ECG feature vector matches. The proposed method is validated by experiments on 113 subjects recorded over wide time duration. It is shown that the PAR parameters can be used to generate different set of features with similar high authentication performance in a doubly secure fashion.