Transcending conventional biometry frontiers: Diffusive Dynamics PPG Biometry

TL;DR

This paper introduces a novel PPG-based biometric authentication method leveraging diffusive dynamics and geometric patterns, offering high stability, anti-spoofing, and suitability for wearable devices.

Contribution

It presents a new biometric approach using diffusive dynamics of PPG signals and geometric pattern analysis, enhancing robustness and anti-spoofing capabilities.

Findings

Achieves very low Equal Error Rates with a single attempt

Demonstrates high stability of biometric features over time

Suitable for integration into wearable biometric systems

Abstract

In the first half of the 20th century, a first pulse oximeter was available to measure blood flow changes in the peripheral vascular net. However, it was not until recent times the PhotoPlethysmoGraphic (PPG) signal used to monitor many physiological parameters in clinical environments. Over the last decade, its use has extended to the area of biometrics, with different methods that allow the extraction of characteristic features of each individual from the PPG signal morphology, highly varying with time and the physical states of the subject. In this paper, we present a novel PPG-based biometric authentication system based on convolutional neural networks. Contrary to previous approaches, our method extracts the PPG signal's biometric characteristics from its diffusive dynamics, characterized by geometric patterns image in the (p, q)-planes specific to the 0-1 test. The diffusive dynamics of the PPG signal are strongly dependent on the vascular bed's biostructure, which is unique to each individual, and highly stable over time and other psychosomatic conditions. Besides its robustness, our biometric method is anti-spoofing, given the convoluted nature of the blood network. Our biometric authentication system reaches very low Equal Error Rates (ERRs) with a single attempt, making it possible, by the very nature of the envisaged solution, to implement it in miniature components easily integrated into wearable biometric systems.