EEG-based emotion recognition using nonlinear feature

Emotions are ubiquitous components of everyday life, as they influence behavior to a large extent. And Emotion recognition is one of the most important and necessary parts in the field of emotion research. Its accuracy relies heavily on the ability to generate representative features. However, this...

Full description

Saved in:
Bibliographic Details
Published in2017 IEEE 8th International Conference on Awareness Science and Technology (iCAST) pp. 55 - 59
Main Authors Jingjing Tong, Shuang Liu, Yufeng Ke, Bin Gu, Feng He, Baikun Wan, Dong Ming
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.11.2017
Subjects
Correlation
correlation dimension
EEG
Electroencephalography
Emotion recognition
Entropy
Feature extraction
Frequency-domain analysis
power spectrum entropy
Support vector machines
Online AccessGet full text

Cover

More Information
Summary:Emotions are ubiquitous components of everyday life, as they influence behavior to a large extent. And Emotion recognition is one of the most important and necessary parts in the field of emotion research. Its accuracy relies heavily on the ability to generate representative features. However, this is a very challenging problem. In this study, EEG nonlinear features, power spectrum entropy and correlation dimension, were extracted to differentiate emotions. International Affective Picture System (IAPS) pictures with different valence but similar arousal level were used to induce the emotions with 8 valence levels. The results showed that the valence levels were positively correlated with these two features, especially in the frontal lobe. Based on the two features, SVM gave an average accuracy of 82.22%. Analyzing the nonlinear features of EEGs is an efficient way to classify emotions.
ISSN:2325-5994
DOI:10.1109/ICAwST.2017.8256518