Emotion recognition by physiological signals

Recently, User’s effective state automatic recognition has become a popular research area. It has many applications ranging from health, education, and personalization. In this paper, emotional state arousal and valence induced by watching video clips are identified by physiological and electroencephalogram (EEG) signals by . After each clip subjects had to assess their feelings about the clip. After doing the first part of data analysis we got robust correlations between users’ self-assessments of arousal and valence. EEG observations were used to train the classifiers for valence recognition and electrocardiogram ECG observations were used for arousal recognition respectively. We achieved averaged results of 71.6% for valence classification for two states and 54.0% for arousal classification for three states.