The inevitable entry of Brain-Computer Interfaces into the market has opened the way for many new opportunities and challenges to be explored. One such opportunity is the use of electroencephalography (EEG) devices for biometric authentication. This study focused on solving the scalability problem of an enterprise authentication system by developing a flexible multi-model system that could handle the addition of extra participants without retraining the entire system. The solution combines the use of a convolutional autoencoder to generalise features from the EEG signal of 20 participants and then feed the encoded features to a one-vs-all SVM classifier for identification.
For the purpose of the study we used a dataset publicly available on IEEE called EEG DATASET OF 7-DAY MOTOR IMAGERY BCI.
A more detailed description of the research methods and the results of the study can be found in the (unpublished) paper.