Person Authentication Using Brain Waves Evoked by Individual-related and Imperceptible Visual Stimuli | IEEE Conference Publication | IEEE Xplore

Person Authentication Using Brain Waves Evoked by Individual-related and Imperceptible Visual Stimuli


Abstract:

Person authentication using biometric information has become increasingly popular owing to the fact that continuous authentication is required in most user management sys...Show More

Abstract:

Person authentication using biometric information has become increasingly popular owing to the fact that continuous authentication is required in most user management systems. This study introduces an individual-related stimulus, instead of common stimuli, to improve the verification performance based on biometric authentication using brain waves evoked by imperceptible visual stimulation. Imperceptible visual stimulation is considered over visual stimulation to overcome obstacles that a user may face when using a system. Compared with previous studies that used circular figures as common stimuli, herein, we ensured a higher evoked response by using individual face image stimulation. Imperceptible stimuli were confirmed by changing the image intensity and presenting a high-speed stimulation. Individual imperceptible face image stimulation confirmed that the following event-related potential (ERP) components: N 170, N 250, and N 400 were obtained. Furthermore, by using various time zones, including the ERP components as features, we verified the performance of eight subjects and achieved an equal error rate (EER) of 6.2%.
Date of Conference: 14-16 September 2022
Date Added to IEEE Xplore: 27 September 2022
ISBN Information:
Electronic ISSN: 1617-5468
Conference Location: Darmstadt, Germany

I. Introduction

Biometric recognition systems can verify the identities of people based on their physiological and/or behavioral personal characteristics, and hence, such systems are widely used as essential safety precautions in various fields such as the military, organizations, universities, banks, airports, as well as in e-commerce. Usually, biometric traits are extracted from facial images, fingerprints, iris images, and palm prints; however, such traits may be easy to steal and copy [1]–[3]. Considering that it is possible to forge authentication systems using stolen data by performing spoofing attacks, it is worthwhile to explore more unique biological signals for biometric authentication purposes. However, several authentication methods do not support the replacement of users in user management systems after authentication considering that users are authenticated only once before system use. Therefore, conventional biometric traits are only used in the begining of user verification process, and hence, such traits may be unsuitable for continuous authentication [4].

References

References is not available for this document.