Abstract:

BACKGROUND: Electroencephalogram (EEG) is a useful tool for clinical diagnosis as well as a good tool for biometric authentication.
OBJECTIVE: To introduce a new method for the identification of a person， by analyzing EEG signals. 
METHODS: Several signal processing methods were used for feature extraction, such as autoregressive model, phase synchronization, energy spectral density, linear complexity, and different methods, including BP neural network, were used for classification of person identification.
RESULTS AND CONCLUSION: The identification rate was more than 80%. The EEG signals of people after stimulation or order performance were recorded, and EEG space-time changes were classified using signal analysis and classification. EEG signals can be used for person identification and generally used in near future.