(WKSP) On the Potential of Data Extraction by Detecting Unaware Facial Recognition with Brain-Computer Interfaces

Abstract

Consumer-grade brain-computer interfaces are becoming more readily available to consumers. Directly reading biological information opens the door for an individual to unwillingly expose personal information. Attackers may be able to glean private information based on the level of recognition a victim has to a specific face, and use that to their advantage. In this work, we use a variety of classification algorithms to classify two types of facial recognition: unaware and aware. To do this, source data is manipulated into two datasets for classification: A set of combined and averaged EEG data, and a set of combined EEG data. We find that in all cases, the combined dataset outperforms the combined and averaged dataset. Further, based on the promising results obtained, there's a risk that a malicious third party could utilize similar techniques to extract private information from individuals without their consent using brain-computer interfaces.