Biometric System Based on EEG Signals: A Nonlinear Model Approach

A research on biometry based on motor imagery EEG signals was described. In this study, I select EEG signals related to motor imagery, and an ARMA model was built. Estimated model parameters vectors as feature vector were extracted, and then to classified by artificial neural networks. Two different classify cases, including authentication and identification, were investigated. Four types of motor imagery EEG signals and three subjects were compared. Experiment results show that EEG carrying individual-specific information which can be successfully exploited for purpose of person authentication and identification.