Recognition of visually induced emotions based on electroencephalography

The aim of this paper is to recognize human emotions using electroencephalography (EEG). Three categories of emotions have been examined: negative, positive and neutral. The emotions were evoked by stimuli in form of slideshows displayed on a computer screen. Electroencephalographic signals were recorded using g.USBAmp with 16 active electrodes. Preprocessing of the signal was done using a common average reference filter. Spectral lines of the signal were used as features. The obtained classification results, for nearest neighbour classifier, proved that distinguishing emotions was possible with good accuracy.