A NEW TIME-FREQUENCY METHOD FOR EEG ARTIFACTS REMOVING

Previous studies have proved the reliability of using EEG signal as a biometric modality. Nowadays, some portable EEG recording systems are emerged as the peripheral devices to allow people use their EEG to play computer games or control toys, and it was also demonstrated that the single-channel EEG signals recorded by the portable equipment can be used for personal identification. However, unlike multi-electrodes devices for medical use, this kind of portable device will amplify noises introducing by power line interference, poor connection of electrode and eyeball movements. So the key point becomes how to effectively remove artifacts and maximally preserve neural activity to reduce adverse effects on EEG identification system. In this paper, a new time-frequency EEG artifacts removing method is proposed. Experimental results show the better performance of the proposed one comparing with the common used EEMD-ICA method.