Electroencephalogram Emotion Recognition Using Variational Modal Decomposition Based Dispersion Entropy Feature Extraction

Electroencephalogram (EEG) emotion recognition has gained considerable attention due to its ability to reflect people's inner emotional states objectively and naturally. Feature extraction is a critical step in EEG emotion recognition because of non-stationarity and irregularity of EEG signals....

Full description

Saved in:
Bibliographic Details
Published in2021 40th Chinese Control Conference (CCC) pp. 3323 - 3326
Main Authors Hu, Si-Jun, Liu, Zhen-Tao, Ding, Xue-Wen
Format Conference Proceeding
LanguageEnglish
Published Technical Committee on Control Theory, Chinese Association of Automation 26.07.2021
Subjects
Brain modeling
Dispersion Entropy
EEG Emotion recognition
Electroencephalography
Emotion recognition
Feature extraction
Real-time systems
Support vector machines
SVM
Temporal lobe
VMD
Online AccessGet full text

Cover

More Information
Summary:Electroencephalogram (EEG) emotion recognition has gained considerable attention due to its ability to reflect people's inner emotional states objectively and naturally. Feature extraction is a critical step in EEG emotion recognition because of non-stationarity and irregularity of EEG signals. A feature extraction method using Variational Modal Decomposition (VMD) to extract Dispersion Entropy (DispEn) is proposed in this paper. Raw EEG signal is decomposed into several components, and DispEn of each component is extracted in eight emotion-related channels. Our method was tested on DEAP dataset in which the EEG emotional states are accessed in Valence-Arousal emotional space. Four emotional states (i.e., HVHA, HVLA, LVHA, LVLA) are classified by Support Vector Machine (SVM). The experimental results show that the accuracy of emotion recognition is 77.87%, which demonstrates its effectiveness.
ISSN:2161-2927
DOI:10.23919/CCC52363.2021.9549884