As a biometric modality for person recognition, electroencephalogram (EEG) has many benefits compared to other biometric modalities. However, studies show that EEG signals are highly affected by the mental task the participant is engaging at the time of EEG recording, as well as being quite sensitive to the applied montage method and the choice of reference point, and also being quite sensitive to the selected segment length. The main goal of this research is to find EEG robust channel set that are least affected by the change of mental task and can be used in biometric applications. Also, since there is so far no consensus on the used montage method or reference point for the EEG recording, so the effect of the EEG montage method or reference point on the selected robust channel set and its recognition performance was analysed. Finally, the effect of the EEG signal segment length on the selected robust channel set and the resulting performance was also analysed. The contribution of this research was to develop a method to analyze stability of the EEG channels and select the robust channel set which will be used in biometric application regardless of the user mental task, also to suggest the best montage method and segment length to be used with this robust channel set. The final outcomes was to propose an optimal EEG channel set, montage method or reference point, and a segment length for the EEG signal for the purpose of biometric applications.
|Date of Award||2017|
|Supervisor||Girija Chetty (Supervisor), Dat Tran (Supervisor) & Wanli Ma (Supervisor)|