Single-channel EEG-based subject identification using visual stimuli

Katsigiannis, Stamos; Arnau-González, Pablo; Arevalillo-Herráez, Miguel; Ramzan, Naeem

Abstract

Electroencephalography (EEG) signals have been recently proposed as a biometrics modality due to some inherent advantages over traditional biometric approaches. In this work, we studied the performance of individual EEG channels for the task of subject identification in the context of EEG-based biometrics using a recently proposed benchmark dataset that contains EEG recordings acquired under various visual and non-visual stimuli using a low-cost consumer-grade EEG device. Results showed that specific EEG electrodes provide consistently higher identification accuracy regardless of the feature and stimuli types used, while features based on the Mel Frequency Cepstral Coefficients (MFCC) provided the highest overall identification accuracy. The detection of consistently well-performing electrodes suggests that a combination of fewer electrodes can potentially provide efficient identification performance, allowing the use of simpler and cheaper EEG devices, thus making EEG biometrics more practical.

Citation

Katsigiannis, S., Arnau-González, P., Arevalillo-Herráez, M., & Ramzan, N. (2021, July). Single-channel EEG-based subject identification using visual stimuli. Presented at 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI), Online

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)
• Start Date: Jul 27, 2021
• End Date: Jul 30, 2021
• Acceptance Date: Jun 8, 2021
• Online Publication Date: Aug 10, 2021
• Publication Date: 2021
• Deposit Date: Jun 8, 2021
• Publicly Available Date: Nov 8, 2021
• Publisher: Institute of Electrical and Electronics Engineers
• DOI: https://doi.org/10.1109/bhi50953.2021.9508581
• Public URL: https://durham-repository.worktribe.com/output/1138822