On the robustness of EEG biometric classification algorithms

Error message

Notice: Undefined offset: 1 in counter_get_browser() (line 76 of /var/www/vhosts/costic1206/public/sites/all/modules/counter/counter.lib.inc).
On the robustness of EEG biometric classification algorithms
Prof Marcos Faundez-Zanuy
Applicant Institution: 
Signal Processing Group - Tecnocampus Mataro-Maresme - Escola Universitaria Politecnica de Mataro - Spain
Host Person: 
Prof Patrizio Campisi
Host Institution: 
Laboratory of Biometrics and Multimedia Forensics - Section of Applied Electronics - Department of Engineering - Università degli Studi Roma Tre - Italy

In this short term mission some classifiers and feature normalization techniques have been applied to EEG signals for biometric recognition. The expertise of the sending institution in biometric recognition of online signature techniques has been applied to EEG signals. Basically direct version comparison with factional distances and vector component weighting have been tested as well as several classifiers. The most interesting approach has been the application of normalization strategies based on cohort normalization. Although this normalization was proposed in the 90’s relevant and recent publications on this line exist. Nevertheless they have never been applied to EEG and this implies a novel research line. Preliminary results are quite encouraging with a reduction in EER by a factor of two.

Grant Period: 
4th Grant Period