Abstract
EEG-based authentication has gained much interest in recent years. However, despite its growing appeal, there are still various challenges to their practical use, such as lack of universality, lack of privacy-preserving, and lack of ease of use. In this paper, we have tried to provide a model for EEG-based authentication by focusing on these three challenges. The proposed method, employing deep learning methods, can capture the fingerprint of the users’ EEG signals for authentication aim. It is capable of verifying any claimed identity just by having a genuine EEG fingerprint and taking a new EEG sample of the user who has claimed the identity, even those who were not observed during the training. The role of the fingerprint function is similar to the hash functions in password-based authentication and it helps preserve the user’s privacy by storing the fingerprint, rather than the raw EEG signals. Moreover, for targeting the lack of ease of use challenge, Gram-Schmidt orthogonalization process reduces the required number of channels to just three ones. The experiments show that the proposed method can reach around 98% accuracy in the authentication of completely new users with only three channels of Oz, T7, and Cz.
Highlights
EEG-based authentication has gained much interest in recent years
We propose a novel and universal EEG biometric authentication system based on the deep learning approach in this research
By combining Convolutional Neural Networks” (CNN) and “Long Short-Term Memory” (LSTM), Sun et al.[14] proposed a model named 1D-Convolutional-LSTM and show that the proposed model could increase the accuracy over a pure CNN model
Summary
Many works have been proposed in the papers on the EEG biometric methods so far, including both shallow and deep methods; though, the shallow classifiers are more common in literature[3]. Just the REO EEG records are used which is already shown that can achieve higher accuracy in subject identification[32,33] As it has been mentioned before, the records are first normalized and split into input data for the model using the sliding window with T = 160 (i.e. one-second segment of EEG signal ) and δ = 4. As it can be seen, the proposed model performs better than the previous works, even with 32 channels and a total of 109 subjects, and achieves the highest possible accuracy As it was mentioned in “Problem statement and motivation” section, an important step to make the EEG authentication more practical is reducing the number of the required channels. The search procedure was stopped because of achieving an acceptable accuracy
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
