CamPressID

Optimizing Camera Configuration and Finger Pressure for Biometric Authentication

Conference paper (2022)

Authors

Weizheng Wang Embedded Systems - EEMCS
Marek Vette Student
Q. Wang Embedded Systems - EEMCS
J. Yang Web Information Systems - EEMCS
Marco Zuniga Embedded Systems - EEMCS
Research Group
Embedded Systems (EEMCS) (TU Delft)
Copyright: © 2022 Weizheng Wang, Marek Vette, Q. Wang, J. Yang, Marco Zuniga
DOI
help_outline
https://doi.org/10.1109/MASS56207.2022.00037
More Info
expand_more

Published Date

2022

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Department

Software Technology

Research Group

Embedded Systems

Abstract

To protect sensitive information on smartphones, state-of-the-art (SoA) studies exploit the built-in camera to capture PPG signals from fingertips as a hard-to-forge biometric. However, those studies do not provide a comprehensive analysis to optimize the camera parameters and finger pressure, leading to distorted and unstable PPG signals that degrade the authentication performance. To overcome these limitations, we propose the CamPressID framework. First, we analyze various camera parameters and optimize their configuration to obtain PPG signals with a high signal-to-noise ratio. Second, we investigate different finger pressures to identify the best pressure for every subject, in order to avoid signal distortion. To evaluate the performance of CamPressID, we collect a diverse dataset with 58 subjects. Our evaluation results show that CamPressID can improve the average balanced accuracy (BAC) by 10%. Moreover, the BAC reaches 90%, which is similar to the accuracy reported in the SoA using a dedicated PPG sensor for authentication.