Research On Emotion Recognition Based On Portable Single-channel EEG Acquisition System

Emotion recognition is to identify people's emotional state by using information analysis and processing method.EEG signals have gradually become one of the most reliable physiological signals to study human emotions.The research of emotion recognition technology based on EEG can promote the development of human-computer interaction,product design,medical care and other fields.It has very important social value and application value.With the rapid development of semiconductor field and dry electrode field,portable EEG acquisition equipment has been developed more mature.The combination of portable EEG acquisition device and emotional recognition technology can rapidly promote the popularization and application of emotional recognition technology,which has important research significance.In this paper,a portable single-channel EEG acquisition system based on TGAM module is built according to the design principle of easy to carry and operate.The system realizes the functions of acquisition,display and storage of EEG signals.In the pretreatment process of single channel EEG signal,an automatic electrooculogram(EOG)artifact removal algorithm combining empirical wavelet transform(EWT),canonical correlation analysis(CCA)and Rayleigh entropy(RE)is proposed to remove the EOG artifacts in single channel EEG quickly and effectively,so as to meet the needs of portable single channel brain-computer interface.In this paper,we designed an experiment to stimulate subjects to produce three different emotions: positive emotions,neutral emotions and negative emotions by using video stimulus materials,and obtained EEG signals of different people under different emotions for emotional recognition.The wavelet energy,wavelet entropy and multi-scale entropy(MSE)of EEG signals are analyzed.A feature extraction algorithm combining discrete wavelet transform(DWT),MSE and principal component analysis(PCA)is proposed to characterize the emotional information contained in EEG signals.In this paper,k-nearest neighbor algorithm(KNN),radial basis function(RBF)neural network classification algorithm and support vector machine(SVM)classification algorithm are studied,and the eigenvalues are put into three classification algorithms.The results show that compared with KNN and RBF algorithm,support vector machine(GA-SVM)classification algorithm based on ant colony optimization has the highest recognition accuracy,which is 75.42%.Finally,in order to further verify the effectiveness of the self-collected data sets and the proposed algorithm,this paper makes a comparative analysis experiment between the open emotional EEG database DEAP and the self-collected data sets.In this paper,a portable single channel EEG acquisition system is built to collect emotional related EEG signals,an automatic EOG artifact removal algorithm for single channel EEG and a feature extraction algorithm combining DWT and MSE are proposed.At the same time,this paper establishes an emotional recognition model to classify emotions.The experimental results show that the designed portable single channel EEG acquisition system can accurately classify and recognize emotions.The research results have certain reference value for the follow-up practical application.