In this paper the use of brain waves as a biometric identifier is investigated. Among the very different protocols that can be used to acquire the electroencephalogram signal (EEG) of an individual we rely on a very simple one: closed eyes in resting conditions. A database of 48 healthy subjects, collected by the authors at the neurophysiology laboratory of the IRCCS Fondazione Santa Lucia, Roma, Italy, has been used for the experiments. Signals acquired from triplets of electrodes have been employed in the experimentations. In more detail, ten different triplets have been used separately in the experiments in order to speculate about the most suitable triplet to capture the occurring phenomena. Feature vectors constituted by the reflection coefficients of a six order AR model have been extracted for each used channel thus giving rise to a feature vector of length eighteen. A polynomial regression based classification is then employed. This analysis has been performed for three different frequency bands for each of the ten different triplet under analysis. The obtained genuine acceptance rate is of 96.08%.
Campisi, P., Scarano, G., Babiloni, F., Fallani, F.D.V., Colonnese, S., Maiorana, E., et al. (2011). Brain waves based user recognition using the Eyes Closed Resting Conditions protocol. In Proceedings of the IEEE International Workshop on Information Forensics and Security (WIFS), 2011 (pp.1-6) [10.1109/WIFS.2011.6123138].
Brain waves based user recognition using the Eyes Closed Resting Conditions protocol
CAMPISI, PATRIZIO;MAIORANA, EMANUELE;
2011-01-01
Abstract
In this paper the use of brain waves as a biometric identifier is investigated. Among the very different protocols that can be used to acquire the electroencephalogram signal (EEG) of an individual we rely on a very simple one: closed eyes in resting conditions. A database of 48 healthy subjects, collected by the authors at the neurophysiology laboratory of the IRCCS Fondazione Santa Lucia, Roma, Italy, has been used for the experiments. Signals acquired from triplets of electrodes have been employed in the experimentations. In more detail, ten different triplets have been used separately in the experiments in order to speculate about the most suitable triplet to capture the occurring phenomena. Feature vectors constituted by the reflection coefficients of a six order AR model have been extracted for each used channel thus giving rise to a feature vector of length eighteen. A polynomial regression based classification is then employed. This analysis has been performed for three different frequency bands for each of the ten different triplet under analysis. The obtained genuine acceptance rate is of 96.08%.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
