Biometric System Based on EEG Signals by Feature Combination

This paper focuses on the person authentication problem based on EEG signals. Many features extracted from EEG recordings have proved to be unique enough between subjects for biometric application. However, different features show different discriminative power for different subjects. Here an authentication system was presented, which makes use of feature combination architecture. The results demonstrate to improve the system performance. Depending on the security level, different thresholds can be applied. The influence of different thresholds on system performance is discussed.