EEG-induced Fear-type Emotion Classification Through Wavelet Packet Decomposition, Wavelet Entropy, and SVM

Authors : Çağlar UYULAN, Ahmet Ergun GÜMÜŞ, Zozan GÜLEKEN

Pages : 241-251

Doi:10.17350/HJSE19030000277

View : 14 | Download : 9

Publication Date : 2022-12-31

Article Type : Research Paper

Abstract :Among the most significant characteristics of human beings is their ability to feel emotions. In recent years, human-machine interface insert ignore into journalissuearticles values(HM); research has centered on ways to empower the classification of emotions. Mainly, human-computer interaction insert ignore into journalissuearticles values(HCI); research concentrates on methods that enable computers to reveal the emotional states of humans. In this research, an emotion detection system based on visual IAPPS pictures through EMOTIV EPOC EEG signals was proposed. We employed EEG signals acquired from channels insert ignore into journalissuearticles values(AF3, F7, F3, FC5, T7, P7, O1, O2, P8, T8, FC6, F4, F8, AF4); for individuals in a visual induced setting insert ignore into journalissuearticles values(IAPS fear and neutral aroused pictures);. The wavelet packet transform insert ignore into journalissuearticles values(WPT); combined with the wavelet entropy algorithm was applied to the EEG signals. The entropy values were extracted for every two classes. Finally, these feature matrices were fed into the SVM insert ignore into journalissuearticles values(Support Vector Machine); type classifier to generate the classification model. Also, we evaluated the proposed algorithm as area under the ROC insert ignore into journalissuearticles values(Receiver Operating Characteristic); curve, or simply AUC insert ignore into journalissuearticles values(Area under the curve); was utilized as an alternative single-number measure. Overall classification accuracy was obtained at 91.0%. For classification, the AUC value given for SVM was 0.97. The calculations confirmed that the proposed approaches are successful for the detection of the emotion of fear stimuli via EMOTIV EPOC EEG signals and that the accuracy of the classification is acceptable.
Keywords : EMOTIV EPOC EEG, Fear emotion, Wavelet Entropy, SVM, ROC