Multi feature fusion EEG emotion recognition

Abstract

In recent years, the research on emotion recognition of EEG signals has attracted much attention. It is an important task to realize the advanced stage of artificial intelligence. How to realize real-time and efficient human-computer interaction has become an important direction of EEG signal research. This study aims to improve the accuracy of emotion recognition of EEG signals, and proposes a binary classification a nd emotion EEG recognition method based on feature fusion is carried out after multi feature extraction to improve the recognition rate. For the preprocessed EEG signals, the eigenvalues extracted from time-frequency, spatial domain, nonlinear dynamics and convolution neural network are used as the initial eigenvectors, and the dimension is reduced by principal component analysis. Finally, the long short-term memory neural network is used for classification. T he emotion recognition experiment w as carried out on the EEG emotion data set deap. The accuracy of two classification i n pleasure a nd arousal w as 8 4.42% a nd 85.61% respectively. The recognition rate is higher than that under single feature and other combined features. The experimental results show that compared with single feature extraction, multi feature fusion has better characteristics of emotional EEG signals, and high classification accuracy c an b e achieved b y using t he long short-term memory neural network. The performance of the emotion recognition method of EEG signals proposed in this paper is better than other methods based on traditional artificial design features and SVM or DBM, It is verified that the method proposed in this paper is feasible.