Photoplethysmography as a form of biometric authentication

This paper explores the feasibility of using signals acquired from a photoplethysmogram (PPG) sensor, typically used in mobile devices for heart rate monitoring, as a valid input signal for biometric authentication systems. Signals were acquired from 10 individuals, generating 708 independent datasets using the Texas Instruments AFE4490/4400 integrated analog front end for pulse oximeters. Results indicate successful identification of individuals with a false acceptance rate of 4.2% and a false rejection rate of 3.7%. Because PPG signal morphology is dependent on the physical state of an individual, biometric authentication cannot occur under stress or coercion which introduces a deviation in heart rate. This technique has demonstrated to be a feasible and simpler replacement to upcoming biometric advancements such as electrocardiogram-based authentication systems, but dependence on consistent measuring conditions limits the capabilities in mobile systems.